G*Power for Sample Size Calculation: A Guide for Biomedical Research

CLYTE research team
1 day ago
5 min read

In the high-stakes world of biomedical and pre-clinical research, determining the correct sample size is not just a statistical formality—it is an ethical and economic imperative. An underpowered study exposes animal subjects to unnecessary risk without yielding reliable data, while an overpowered study wastes valuable resources and time.

This guide provides a detailed summary of how to use G*Power, the gold-standard free software for power analysis, tailored specifically for researchers analyzing biomedical data.

You might also be interested in this t-test vs Anova article!

Ask our Soφ AI about G*Power calculation!

Why Sample Size Matters in Pre-Clinical Research

According to recent guidelines published in The Journal of Educational Evaluation for Health Professions and archived in PubMed Central, scientifically rigorous research relies on accurate estimates of therapeutic effects.

Scientific Validity: Studies with inadequate power cannot reliably detect significant differences, leading to Type II errors (false negatives).
Ethical Responsibility: It is unethical to conduct animal or human trials if the study design cannot statistically answer the research question.
Economic Efficiency: Proper calculation prevents the waste of expensive reagents, drug compounds, and laboratory hours.

Key Concepts: The "Input Parameters"

Before opening G*Power, every researcher must define four key parameters. Understanding these is crucial for accurate biomedical data analysis.

1. Alpha (ɑ) – The Significance Level

This is the probability of a Type I error (false positive). In most biomedical fields, this is standardly set to 0.05 (5%), meaning there is a 5% risk of rejecting the null hypothesis when it is actually true.

2. Power (1-β)

Power is the probability of correctly rejecting the null hypothesis when it is false (detecting a true effect).

Standard Practice: A power of 0.80 (80%) is the widely accepted minimum. This implies a 20% risk of a Type II error (false negative).
Pre-Clinical Context: For critical toxicity or efficacy studies, researchers may increase this to 0.90 or 0.95.

3. Effect Size

This is the magnitude of the difference you expect to see (e.g., the difference in tumor size between treated and control mice).

Cohen’s Conventions: G*Power offers default values (Small: 0.2, Medium: 0.5, Large: 0.8 for t-tests).
Best Practice: Ideally, estimate this from pilot data or literature review rather than relying solely on defaults.

4. Sample Allocation Ratio

This defines the balance between groups. An allocation ratio of 1 indicates equal group sizes (N_1 = N_2), which generally provides the most statistical power.

Step-by-Step Tutorial: The Independent t-Test

A common scenario in pre-clinical research is comparing the efficacy of a new drug (Group A) versus a vehicle control (Group B). Here is how to calculate the sample size using G*Power.

Step 1: Select the Test Family

Launch G*Power.
Under Test Family, select t tests.
Under Statistical Test, select Means: Difference between two independent means (two groups).

Step 2: Choose Power Analysis Type

Select "A priori: Compute required sample size - given ɑ, power, and effect size".
Note: Use "Post hoc" only if you have already finished the study and need to calculate the achieved power.

Step 3: Input Parameters

Enter the values determined by your study design:

Tail(s): Select Two (unless you have a strong justification for a one-sided hypothesis).
Effect size d: Enter your calculated value (e.g., 0.5 for a medium effect).
ɑ err prob: 0.05.
Power (1-β err prob): 0.80.
Allocation ratio N2/N1: 1.

Step 4: Calculate and Interpret

Click Calculate.
Look at the Output Parameters.
Total Sample Size: This is the minimum total number of subjects required.
Actual Power: G*Power often provides a value slightly higher than 0.80 because sample sizes must be whole numbers.

Example Result: For a medium effect size (ɗ=0.5), ɑ=0.05, and Power=0.80, G*Power will recommend a total sample size of 128 participants (64 per group).

Reporting Your Sample Size Calculation

To ensure your manuscript meets Consort and Arrive guidelines, report your calculation clearly:

"An a priori power analysis was conducted using G*Power (Ver. 3.1.9.7). To test the hypothesis that the new drug treatment differs from the control, a sample size of 64 subjects per group was determined to be sufficient to achieve 80% power with an ɑ of 0.05, assuming a medium effect size (ɗ=0.5). The total required sample size was 128."

Conclusion

Mastering G*Power is an essential skill for modern biomedical researchers. By accurately calculating sample sizes, you ensure your pre-clinical data is robust, reproducible, and ready for publication in high-impact journals.

More step-by-step protocols like this?!

Frequently Asked Questions (FAQ) about G*Power for Sample Size Calculation

How to use G*Power for sample size?

To use G*Power for sample size determination, follow these four basic steps:

Select Test Family: Choose the statistical test you plan to use (e.g., t-tests for comparing two groups, F-tests for ANOVA).
Choose Analysis Type: Select "A priori" to calculate the required sample size before starting your study.
Input Parameters: Enter your required significance level (ɑ, usually 0.05), desired Power (usually 0.80), and estimated Effect Size (derived from pilot data or literature).
Calculate: Press the "Calculate" button to generate the minimum total sample size required for your biomedical study.

How to calculate sample size based on power?

In G*Power, calculating sample size based on power is done using an A Priori Power Analysis. Instead of guessing the sample size, you input your desired Power level (typically 0.80 or 80%) along with your significance level (ɑ) and effect size. The software then mathematically determines the specific number of subjects (N) needed to achieve that probability of detecting a true effect, ensuring your study is not underpowered.

What does 80% power in a study mean?

80% power means that if a true effect or difference exists in your experiment (e.g., a drug actually works), your study has an 80% probability of detecting it and rejecting the null hypothesis. Conversely, it implies a 20% risk of committing a Type II error (false negative), where you fail to detect an effect that is actually present. This is the standard acceptable threshold in most pre-clinical and clinical research.

What is the formula for calculating sample size?

While G*Power automates complex calculations for various distributions (t, F, x^2), the fundamental manual formula for comparing two means (independent t-test) helps illustrate the relationship between variables:

n = [2σ^2 (Z_ɑ/2 + Z_β)^2 ] / △^2

Where:

n = Sample size per group
σ = Standard deviation (variance)
Z_ɑ/2 = Z-score for the significance level (e.g., 1.96 for ɑ=0.05)
Z_β = Z-score for the desired power (e.g., 0.84 for 80% power)
△ = The expected mean difference (effect size)

What is G*Power known for?

G*Power is a free, open-source statistical software tool developed by researchers at the Heinrich Heine University Düsseldorf. It is globally recognized as the standard tool for conducting power analyses and sample size calculations in social, behavioral, and biomedical sciences. It is favored for its ability to handle a wide variety of statistical tests (t-tests, ANOVA, regression, correlation) and its ease of use for both a priori (planning) and post hoc (evaluation) analyses.