Math Calculators

Sample Size Calculator

Calculate the minimum sample size needed for statistically significant results in research, surveys, and A/B testing with customizable confidence levels (90%, 95%, 99%) and margin of error settings. Essential for researchers, marketers, data scientists, and product managers to determine adequate sample sizes for reliable statistical analysis and hypothesis testing.

How to Use the Sample Size Calculator

Use the Sample Size Calculator to the minimum sample size needed for statistically significant results in research, surveys, and A/B testing with customizable confidence levels (90%, 95%, 99%) and margin of error settings. Essential for researchers, marketers, data scientists, and product managers to determine adequate sample sizes for reliable statistical analysis and hypothesis testing.. Enter your values to get accurate, instant results tailored to your situation.

Free math calculators for algebra, geometry, statistics, and more. Solve complex mathematical problems with step-by-step solutions.

Common Uses

• Calculate sample size accurately and instantly
• Calculate statistical significance accurately and instantly
• Calculate research accurately and instantly
• Calculate statistics accurately and instantly
• Calculate confidence interval accurately and instantly

Related Calculators

• Statistics Calculator
• Percentage Calculator
• Z-Score Calculator

More Math Calculators

• Percentage Calculator
• Scientific Calculator
• Base Converter
• Factorial Calculator
• Greatest Common Divisor Calculator
• Least Common Multiple Calculator

Browse all 311+ free online calculators

Sample Size Guide

Statistical sampling

Expert Tips

• Larger Sample = More Confidence, Smaller Margin of Error: 95% confidence, ±5% margin, 50% variance: n=385. 95% confidence, ±3% margin: n=1,067. 99% confidence, ±3% margin: n=1,849. Doubling confidence or halving margin requires 4× sample size. Diminishing returns on large samples.
• Formula: n = (Z² × p × (1-p)) / E²: Z = confidence z-score (1.96 for 95%, 2.58 for 99%). p = expected proportion (use 0.5 for maximum variance). E = margin of error (0.05 for ±5%). Example: n = (1.96² × 0.5 × 0.5) / 0.05² = 384.16 ≈ 385 samples.
• Population Size Matters Only When <20× Sample: Population 1,000,000 vs 10,000 → same sample size (385 for ±5%). Population 500 → need finite population correction → sample 217 instead of 385. Small populations require smaller samples (but larger %) than infinite populations.
• Random Sampling Critical—Size Doesn't Fix Bias: 10,000 biased samples < 100 random samples. Online polls, convenience samples unreliable regardless of size. Ensure: Random selection, representative sample, no self-selection bias. Bad sampling = bad results at any size.

Essential Fundamentals — Sample size determination

Key Parameters

• Confidence Level: Probability result within margin of error. 90%: Less confident (Z=1.65). 95%: Standard (Z=1.96). 99%: Very confident (Z=2.58). Higher confidence needs larger sample. 95% most common for surveys, 99% for critical research.
• Margin of Error: Acceptable range of error. ±3%: Tight precision, large sample. ±5%: Standard for surveys. ±10%: Rough estimate, small sample. Smaller margin = larger sample (quadratic relationship). Halve margin = 4× sample size.

Advanced Strategies — Sample optimization

Practical Considerations

• Finite Population Correction: For small populations: n_adjusted = n / (1 + (n-1)/N). Population 500, initial n=385 → 217 actual samples needed. Reduces required sample size when population is finite and small.
• Stratified Sampling: Divide population into groups (strata), sample proportionally from each. More precise than simple random sampling. Reduces required total sample size while maintaining accuracy across subgroups.

Frequently Asked Questions

What sample size do I need for my survey?

Sample size factors: Population size: Small (<1,000): Need ~300 for 5% margin. Medium (1K-100K): Need ~370-380 for 5% margin. Large (>100K): Need ~385 (population size matters less). Confidence level (how sure results are accurate): 90%: Z=1.645, smaller sample. 95%: Z=1.96 (most common, balance of cost vs accuracy). 99%: Z=2.576, larger sample (very confident). Margin of error (acceptable error range): ±1%: Need ~9,600 samples (very precise). ±3%: Need ~1,067 samples (good precision). ±5%: Need ~385 samples (acceptable for most surveys). ±10%: Need ~97 samples (rough estimate only). Common scenarios: Quick poll (±10% margin, 90% confidence): ~80 samples. Standard survey (±5% margin, 95% confidence): ~385 samples. Academic research (±3% margin, 99% confidence): ~1,800 samples. Precise measurement (±1% margin, 99% confidence): ~16,600 samples. Response rate: Assume 30% response rate. If need 385 responses, contact 1,283 people (385 ÷ 0.30).

What is confidence level and margin of error?

Confidence level: Definition: Probability that true value falls within margin of error. Example (95% confidence): If 100 surveys run, 95 would have results within margin of error. 5 would miss due to random chance. Higher confidence = need more samples. 90%: Less sure, smaller sample (cost-effective). 95%: Standard (balance of accuracy and cost). 99%: Very sure, larger sample (expensive). Margin of error: Definition: Range around survey result where true value likely falls. Example (±5% margin): Survey says 60% support. True value likely between 55-65%. Smaller margin = more precise, need more samples. ±1%: Very precise (large sample needed, expensive). ±3%: Good precision (academic standard). ±5%: Acceptable (common for public polls). ±10%: Rough estimate (quick/cheap polls). Real-world example: Poll: "55% support candidate" (95% confidence, ±3% margin). Interpretation: 95% confident true support is 52-58%. 5% chance true value outside this range. To improve precision: ±1% margin: Need 3× more samples (expensive). 99% confidence: Need 2× more samples (expensive). Trade-off: Higher confidence + smaller margin = larger sample = higher cost.

How does population size affect sample size?

Population size impact: Small populations (<1,000): Sample = large % of population. Example: 500 population needs ~220 sample (44%). Medium populations (1K-100K): Sample stabilizes around 370-385. Example: 10K population needs ~370 (3.7%). 100K population needs ~383 (0.4%). Large populations (>100K): Sample approaches 385 regardless of size. Example: 1M population needs ~384 (0.04%). 100M population needs ~385 (0.0004%). Why: Small populations: Sampling significant % reduces variability. Large populations: Sampling tiny % still captures diversity. After ~100K, size barely matters (diminishing returns). Formula adjustment: Infinite population: n₀ = (Z² × 0.25) / e². Finite population: n = n₀ / (1 + (n₀-1) / N). As N→∞, n→n₀ (finite correction disappears). Practical examples: University (5K students, ±5% margin, 95%): Need 357 samples (7%). City (100K residents, ±5%, 95%): Need 383 samples (0.4%). Country (300M people, ±5%, 95%): Need 385 samples (0.0001%). Bottom line: For populations >100K, use 385 samples (±5%, 95%) regardless of size. Smaller populations need proportionally more (but <50% ever).