Finite Population Correction Calculator

What Is the Finite Population Correction?

When sampling a substantial portion of a population, standard error formulas overstate uncertainty. The finite population correction (FPC) adjusts for this.

Standard formula assumes: Infinite population (or sampling with replacement) Reality: Sampling 10% of a population is more precise than formulas suggest

The Correction Factor

FPC = √[(N - n) / (N - 1)]

Where:

• N = population size
• n = sample size

Multiply the standard error by FPC:

SE_adjusted = SE_standard × FPC

When FPC Matters

Rule of thumb: Apply FPC when n/N > 0.05 (sampling more than 5% of the population)

When you sample 100% of the population, there's no sampling error—you have the true value.

Practical Examples

Small Employee Survey

Population: N = 200 employees Sample: n = 80 (40% of population)

FPC = √[(200-80)/(200-1)] = √[120/199] = 0.777

Standard error is 22% smaller than the infinite-population formula suggests.

Customer Satisfaction

Population: N = 1,000 customers Sample: n = 400 (40% of population)

FPC = √[(1000-400)/(1000-1)] = √[600/999] = 0.775

Large Population

Population: N = 100,000 Sample: n = 1,000 (1% of population)

FPC = √[(100,000-1,000)/(100,000-1)] = √[99,000/99,999] = 0.995

FPC is negligible—standard formulas work fine.

Effect on Confidence Intervals

Adjusted margin of error:

ME_adjusted = ME_standard × FPC

Example:

• Standard ME = ±5%
• FPC = 0.866
• Adjusted ME = ±5% × 0.866 = ±4.3%

The interval is narrower because sampling much of the population reduces uncertainty.

Effect on Sample Size

For a target margin of error with finite population:

n_adjusted = n_infinite / [1 + (n_infinite - 1)/N]

Example: Target ±5% at 95% confidence

• Infinite population: n = 385
• Population N = 1,000: n = 385 / [1 + 384/1000] = 278

You need fewer respondents from a small population.

Applications

Internal Surveys

Surveying employees, club members, or other known populations:

• Population is enumerable
• Often sample significant fraction
• FPC substantially reduces required sample

Quality Control

Inspecting products from a batch:

• Batch size is known
• May inspect 10-50% of batch
• FPC affects inspection conclusions

Census Accuracy Assessment

Even censuses have nonresponse. Post-enumeration surveys to check accuracy:

• Sample from known population
• FPC applies

Finite Event Studies

Analyzing all occurrences within a timeframe:

• "All customer complaints last month"
• If sampling from this defined set, FPC may apply

When FPC Doesn't Apply

Process sampling: Sampling from an ongoing process (production line, web traffic) where the "population" is theoretically infinite.

Sampling with replacement: If each selection is replaced before the next draw, FPC doesn't apply.

Inference to larger population: If your sample of 100 from 500 employees aims to represent all potential employees, FPC doesn't apply.

Degrees of Freedom

With FPC, degrees of freedom may also need adjustment in some analyses.

For variance estimation: df_adjusted = n - 1 (unchanged, typically)

But effective sample size may be considered larger due to reduced variance.

The Hypergeometric Connection

FPC arises from the hypergeometric distribution (sampling without replacement) versus binomial (sampling with replacement).

As N → ∞, hypergeometric → binomial, and FPC → 1.

Historical Context

The finite population correction has been part of survey methodology since its formalization in the early 20th century.

William Cochran systematically treated FPC in his influential textbook "Sampling Techniques" (1953, 1963, 1977).

The correction is mathematically straightforward but often overlooked in practice, leading to overly conservative intervals and larger-than-necessary samples.

Common Mistakes

Applying to Infinite Processes

Web traffic, ongoing production, or any process that continues indefinitely has no finite N to correct for.

Ignoring When Applicable

Surveys of known, small populations waste resources if FPC isn't used in sample size planning.

Misidentifying the Population

The population is what you're making inferences about, not just who's available to survey.

Related Calculations

• Survey Sample Size Calculator — Sample size with FPC
• Standard Error Calculator — Unadjusted SE
• Margin of Error Calculator — Apply FPC to ME
• Confidence Interval for Mean — Adjust CI width

Summary

The finite population correction reduces standard errors and confidence interval widths when sampling a significant fraction of the population. Apply it when n/N exceeds 5%. The correction factor √[(N-n)/(N-1)] approaches 1 for large populations and 0 for censuses. In small-population surveys—employee surveys, member organizations, finite batches—FPC can substantially reduce required sample sizes and produce tighter confidence intervals that correctly reflect reduced uncertainty from sampling much of the population.