Animal Sample Size Calculator
How many animals per group do you need? Run a power analysis for two means or two proportions, with built-in attrition adjustment and 3Rs guidance.
Try Moustra freeHow sample size and power analysis works
A sample size calculation answers one question: how many subjects do you need so that, if a real difference exists, you will reliably detect it? The answer depends on four quantities.
Alpha (significance level) is the probability of a false positive — concluding there is an effect when there is none. The conventional threshold is 0.05 (two-sided), meaning you accept a 5% false-alarm rate. Stricter studies use 0.01.
Power (1 − beta) is the probability of detecting a real effect — a power of 0.80 means you have an 80% chance of reaching statistical significance if the true effect is at least as large as you specified. Higher power requires more animals.
Effect size captures how different the groups are relative to natural variability. For continuous outcomes, Cohen's d = |mean1 − mean2| / SD. For binary outcomes, the absolute difference between proportions is used. Smaller effect sizes require larger samples.
For two means, the formula is approximately n = 2(z_alpha/2 + z_beta)² / d², where z values are the standard-normal quantiles corresponding to alpha and power. For two proportions, the numerator uses the pooled variance of both rates. In both cases, the result is rounded up to a whole number of animals.
Attrition inflation adjusts upward so that even after expected losses (deaths, exclusions, missing data) you still have enough completers to reach the required power. If you expect 10% dropout, enrol n / (1 − 0.10) animals per group.
How to use this calculator
- Choose your analysis type — select "Compare two means" for continuous outcomes (weight, volume, concentration) or "Compare two proportions" for binary outcomes (survival, response rate).
- Set alpha and power — 0.05 and 0.80 are the conventional defaults. Use 0.90 or 0.95 power when the cost of a false negative is high.
- Enter expected attrition — use your lab's historical loss rate or a conservative estimate (10–20% is common for rodent studies). The calculator inflates enrolment automatically.
- Enter your effect parameters — for means, provide both group means and the expected standard deviation (from pilot data or the literature). For proportions, enter the expected response rate in each group as a percentage.
- Read the results — the highlighted box shows animals to enrol per group and the total. Export a CSV for your IACUC submission or share the URL with your collaborators.
Frequently Asked Questions
How many animals do I need for my experiment?
The number depends on the expected effect size (how different you expect the groups to be), the variability of your outcome, the significance level (alpha), and the statistical power you want. Enter your expected group means (or proportions) and standard deviation into this calculator and it will give you the minimum animals per group and the total to enrol after accounting for expected losses.
What is statistical power and what alpha should I choose?
Statistical power is the probability of detecting a real effect if one exists — 0.80 (80%) is the conventional minimum, meaning a 20% chance of a false negative. Alpha is the false-positive rate; 0.05 is the standard threshold, meaning you accept a 5% chance of a significant result when there is actually no effect. Lower alpha or higher power both require more animals.
How do I choose an effect size or Cohen's d?
Cohen's d is the difference between the two group means divided by the pooled standard deviation. If you have pilot data or literature values, enter the group means and SD directly and the calculator derives d for you. As a rough guide: d = 0.2 is small, 0.5 is medium, and 0.8 is large, but for animal studies a biologically meaningful difference — not an arbitrary benchmark — should drive the choice.
How does sample size relate to the 3Rs and IACUC justification?
The 3Rs principle of Reduction requires using the minimum number of animals necessary to achieve statistically valid results. IACUC protocols typically require a formal power analysis that shows the chosen sample size is neither wasteful (too many) nor underpowered (too few). This calculator produces the numbers you need for that justification, and Moustra can track your cohorts against the approved numbers throughout the study.
What is the difference between comparing two means and comparing two proportions?
Use the two-means test when your outcome is a continuous measurement — body weight, tumor volume, biomarker concentration, etc. Use the two-proportions test when your outcome is binary — survival, response, presence/absence of a phenotype. The underlying formulas differ: means rely on the t-distribution approximation (z-test for large samples), while proportions use a pooled variance based on both group rates.
More free tools
View all tools →Breeding Genetics Calculator
Predict litter genotype & phenotype ratios and plan breeding pairs for mouse, fly, fish, rat & worm crosses.
Punnett Square Calculator
Type any two parent genotypes (e.g. AaBb × Aa) and instantly see the full Punnett square and offspring ratios.
Mouse Coat Color Calculator
Cross two mice by their A/B/C/D/P loci and see the expected coat colors of the litter, with epistasis handled for you.
Ready to Transform Your Lab Management?
Researchers at top institutions use Moustra to streamline their mouse colony management. Start free today and experience the difference.
Try Moustra For Free