Power Sample Size Calculator
Our biostatistics calculator computes power sample size accurately. Enter measurements for results with formulas and error analysis.
Calculator
Adjust values & calculatePower Curve (n per group vs Power)
Formula
Where n is the sample size per group, z_alpha is the critical z-value for the chosen significance level, z_beta is the z-value corresponding to the desired power (1 - beta), and d is Cohen's d effect size (mean difference divided by pooled standard deviation). For a one-sample test, remove the factor of 2. The formula assumes equal group sizes for two-sample tests.
Last reviewed: December 2025
Worked Examples
Example 1: Clinical Drug Trial Sample Size
Example 2: Gene Expression Study with Small Effect
Background & Theory
The Power Sample Size Calculator applies the following established principles and formulas. Biology is the scientific study of life, encompassing the structure, function, growth, evolution, and distribution of living organisms. At the cellular level, all life is composed of cells, the basic structural and functional units of organisms. Prokaryotic cells lack a membrane-bound nucleus, while eukaryotic cells possess a nucleus and membrane-bound organelles including mitochondria, which generate ATP through oxidative phosphorylation, and ribosomes, which synthesize proteins. Genetics quantifies the inheritance of traits. Gregor Mendel's laws describe how alleles segregate during gamete formation and assort independently for genes on different chromosomes. Punnett squares provide a visual method for calculating the probability of offspring genotypes and phenotypes from known parental genotypes. For a monohybrid cross of two heterozygotes (Aa ร Aa), the expected phenotypic ratio is 3 dominant to 1 recessive. The Hardy-Weinberg equilibrium principle states that allele and genotype frequencies in a population remain constant from generation to generation in the absence of evolutionary forces. If p and q are the frequencies of two alleles at a locus, then p + q = 1 and genotype frequencies are pยฒ, 2pq, and qยฒ for the three possible genotypes. Deviations from equilibrium signal the action of natural selection, genetic drift, mutation, migration, or non-random mating. Population growth follows two primary models. Exponential growth, N = Nโeสณแต, describes unlimited growth where Nโ is the initial population, r is the intrinsic rate of increase, and t is time. Logistic growth incorporates carrying capacity K, describing how growth slows as population approaches the environment's maximum sustainable size: dN/dt = rN(1 โ N/K). Enzyme kinetics describes the rate of enzyme-catalyzed reactions. The Michaelis-Menten equation, v = Vmax[S]/(Km + [S]), relates reaction velocity v to substrate concentration [S], maximum velocity Vmax, and the Michaelis constant Km, which equals the substrate concentration at half-maximal velocity. DNA replication relies on complementary base pairing: adenine pairs with thymine (two hydrogen bonds) and guanine with cytosine (three hydrogen bonds), ensuring faithful copying of genetic information.
History
The history behind the Power Sample Size Calculator traces back through the following developments. The systematic study of living things began with Aristotle (384โ322 BCE), who classified over 500 animal species and wrote foundational texts on anatomy, reproduction, and animal behavior. His scala naturae ranked organisms in a hierarchy from simple to complex and influenced biological thought for two millennia. Theophrastus, his student, applied similar methods to plants. Carl Linnaeus established modern taxonomy in Systema Naturae (1735), introducing the binomial nomenclature system that assigns each organism a genus and species name. His hierarchical classification system โ species, genus, family, order, class, phylum, kingdom โ provided the organizational framework that biologists still use, now extended to seven ranks and supplemented by cladistics. Charles Darwin and Alfred Russel Wallace independently developed the theory of evolution by natural selection, which Darwin published in On the Origin of Species in 1859. Darwin argued that heritable variation exists within populations, that organisms with advantageous traits survive and reproduce at higher rates, and that this differential reproduction gradually changes the character of populations over generations. This unified all of biology under a single explanatory framework. Gregor Mendel's meticulous pea plant experiments, conducted from 1856 to 1863 and published in 1866, established the particulate nature of inheritance and the laws of segregation and independent assortment. Overlooked until 1900, when three botanists independently rediscovered his work, Mendel's laws laid the foundation for the science of genetics. James Watson and Francis Crick, building on Rosalind Franklin's X-ray crystallography data, determined the double-helix structure of DNA in 1953, revealing the physical basis of heredity and the mechanism by which genetic information is stored and copied. The Human Genome Project, a 13-year international collaboration, published the complete sequence of the human genome in 2003, comprising approximately 3.2 billion base pairs. The development of CRISPR-Cas9 gene editing by Jennifer Doudna, Emmanuelle Charpentier, and colleagues from 2012 onward opened an era of precise genome modification with transformative implications for medicine, agriculture, and basic research.
Key Features
- Computes a full descriptive statistics summary from a data set, including mean, median, mode, range, variance, standard deviation, skewness, and interquartile range.
- Constructs confidence intervals for population proportions and means at any confidence level, displaying the margin of error, standard error, and critical value used.
- Calculates p-values and test statistics for z-tests, one- and two-sample t-tests, and chi-square goodness-of-fit and independence tests, with automatic two-tailed or one-tailed selection.
- Performs ordinary least squares linear regression on paired data, returning the slope, intercept, R-squared value, and a residual summary to assess model fit.
- Evaluates the CDF and PDF for major probability distributions including the normal, binomial, and Poisson distributions, given user-supplied parameters and input values.
- Determines the required sample size to achieve a specified margin of error and confidence level for both proportion and mean estimation problems.
- Computes the Pearson and Spearman correlation coefficients between two variables, indicating the strength and direction of their linear or monotonic relationship.
- Applies Bayes' theorem to calculate posterior probabilities given a prior probability, likelihood, and marginal likelihood, with a clear breakdown of each term in the formula.
Frequently Asked Questions
Formula
n = ((z_alpha + z_beta)^2 * 2) / d^2
Where n is the sample size per group, z_alpha is the critical z-value for the chosen significance level, z_beta is the z-value corresponding to the desired power (1 - beta), and d is Cohen's d effect size (mean difference divided by pooled standard deviation). For a one-sample test, remove the factor of 2. The formula assumes equal group sizes for two-sample tests.
Worked Examples
Example 1: Clinical Drug Trial Sample Size
Problem: A researcher plans a two-group clinical trial expecting a medium effect size (d=0.5) and wants 80% power at alpha=0.05 (two-tailed). How many patients per group?
Solution: z_alpha/2 = z(0.025) = 1.96\nz_beta = z(0.20) = 0.842\nn per group = ((1.96 + 0.842)^2 * 2) / 0.5^2\n= (2.802^2 * 2) / 0.25\n= (7.851 * 2) / 0.25\n= 15.702 / 0.25 = 62.8\nRound up: n = 63 per group
Result: 63 patients per group (126 total) needed for 80% power to detect a medium effect
Example 2: Gene Expression Study with Small Effect
Problem: A genomics study expects a small effect (d=0.3) and needs 90% power at alpha=0.01 (two-tailed). Calculate required sample size.
Solution: z_alpha/2 = z(0.005) = 2.576\nz_beta = z(0.10) = 1.282\nn per group = ((2.576 + 1.282)^2 * 2) / 0.3^2\n= (3.858^2 * 2) / 0.09\n= (14.884 * 2) / 0.09\n= 29.768 / 0.09 = 330.8\nRound up: n = 331 per group
Result: 331 samples per group (662 total) needed for 90% power at alpha=0.01 with small effect
Frequently Asked Questions
What is statistical power and why does it matter?
Statistical power is the probability that a test will correctly reject a false null hypothesis (i.e., detect a real effect when one exists). A power of 0.80 means there is an 80% chance of detecting the effect if it truly exists, and a 20% chance of a Type II error (missing the effect). In biological research, underpowered studies waste resources and may fail to detect important effects like drug efficacy or genetic associations. Most journals and regulatory agencies require a minimum power of 0.80, though 0.90 is recommended for critical studies.
How do I choose an appropriate effect size?
Effect size (Cohen's d) quantifies the magnitude of the difference between groups relative to variability. Cohen suggested benchmarks: small (d=0.2), medium (d=0.5), and large (d=0.8). However, it is better to base your effect size on prior research, pilot studies, or the minimum clinically meaningful difference. For example, if a drug must reduce blood pressure by at least 5 mmHg (SD=10) to be clinically relevant, d = 5/10 = 0.5. Using standardized benchmarks without domain knowledge can lead to misleadingly sized studies.
What is the relationship between sample size, power, and effect size?
These three quantities are mathematically linked: increasing any one allows you to decrease another. Larger sample sizes increase power for a given effect size. Larger effect sizes require smaller samples for the same power. Common trade-offs include: to detect a small effect (d=0.2) at 80% power requires about 394 per group; a medium effect (d=0.5) requires about 64 per group; a large effect (d=0.8) requires only about 26 per group. Doubling sample size does not double power; the relationship follows a curve that flattens as power approaches 1.0.
How does the significance level (alpha) affect sample size?
A smaller alpha (e.g., 0.01 vs 0.05) means stricter criteria for significance, requiring a larger sample to achieve the same power. At alpha = 0.05 and power 0.80 for medium effect, you need about 64 per group. At alpha = 0.01, this increases to about 95 per group. In genomics and multiple-testing scenarios, researchers often use Bonferroni-corrected alpha values (e.g., 0.05/1000 = 0.00005), which dramatically increases required sample sizes and is why genome-wide association studies need thousands of participants.
Why might my result differ from another tool or reference?
Differences typically arise from rounding conventions, the specific version of a formula (for example, simple vs compound interest), or unit inconsistencies between inputs. Check that both tools are using the same formula variant and the same units. The References section links to the authoritative source behind the formula used here.
What inputs do I need to use Power Sample Size Calculator accurately?
Each field is labelled with the required unit (metric or imperial). Gather your source values before starting โ for example, a weight measurement in kilograms, a distance in metres, or a dollar amount โ and enter them exactly as measured. The formula section on this page lists every variable and explains what each represents.
References
Reviewed by Daniel Agrici, Founder & Lead Developer ยท Editorial policy