Question 1

Why does cutting the margin of error in half more than double the sample size?

Accepted Answer

Margin of error sits in the denominator of the formula as E², not E — so the relationship is an inverse square, not a straight inverse. Halving your margin of error from ±10% to ±5% doesn't double the required respondents, it quadruples them (from roughly 97 to 385 at 95% confidence). Going from ±5% to ±2.5% quadruples it again to around 1,537. This is why tightening precision from "pretty good" to "very precise" gets expensive fast, and why most consumer surveys settle for ±3% to ±5% rather than chasing ±1%.

Question 2

Why does this calculator default the population proportion to 50%?

Accepted Answer

The p(1−p) term in the formula peaks exactly at p = 0.5 (0.5 × 0.5 = 0.25, the largest value that product can take) and shrinks as the split moves toward either extreme, such as 0.9 × 0.1 = 0.09. Since you almost never know the true proportion before you survey — that's the whole point of running the survey — using 50% gives you the largest, most conservative sample size that guarantees your margin of error holds no matter which way the real answer falls. If prior data or a pilot study gives you a reasonable estimate of the true proportion, plugging that in instead will lower the required sample size.

Question 3

Does surveying a bigger population always mean I need more respondents?

Accepted Answer

No, and this surprises most first-time survey planners. Beyond a certain population size, the required sample size plateaus and barely moves — the finite population correction shows why: n_adjusted = n / (1 + (n−1)/N). Surveying a city of 500,000 versus a country of 300 million needs almost the same number of respondents for the same confidence level and margin of error, because once N is much larger than the uncorrected n, the correction term approaches 1 and stops shrinking anything further. The correction only makes a meaningful difference when your population is small relative to the sample size the base formula asks for — a company of 200 employees, a classroom of 30 students, a client list of 1,000.

Question 4

What's the practical difference between 95% and 99% confidence?

Accepted Answer

The Z-score jumps from 1.96 to 2.576, and because it's squared in the formula, the required sample size increases by roughly 73% for the same margin of error and proportion — a ±5% margin needs about 385 respondents at 95% confidence versus about 664 at 99%. In practice, 95% is the standard for most market research, academic surveys, and business decisions, while 99% is reserved for higher-stakes contexts like clinical trials or regulatory reporting where a wrong conclusion is far more costly than the extra respondents. Once you've collected your data, run it back through our confidence interval calculator, which works in the opposite direction — turning your actual sample results into a margin of error, rather than planning the sample size before you start.

Margin of Error	90% CI	95% CI	99% CI
±1%	6,766	9,604	16,590
±2%	1,692	2,401	4,148
±3%	752	1,068	1,844
±4%	423	601	1,037
±5%	271	385	664
±7%	139	196	339
±10%	68	97	166

Sample Size Calculator