How to Calculate Sample Size? G*Power Guide (2026)
A step-by-step guide on calculating sample size for clinical and academic research studies using G*Power software, with a focus on statistical power and assumptions.
One of the most critical steps in the planning phase of a scientific study is the correct determination of sample size calculation. Studies conducted with insufficient sample sizes may fail to detect statistically significant effects (Type II error), while excessively large samples may lead to unnecessary use of resources and ethical concerns.
For this reason, power analysis has become a standard practice in modern research. One of the most widely used free tools for this purpose is the G*Power software.
Key Components of Power Analysis
Before performing sample size calculation in G*Power, it is essential to understand the following four statistical concepts:
1. Type I Error (α - Alpha)
Commonly set at 0.05. It represents the probability of finding a difference when none actually exists (false positive).
2. Power (1-β)
Typically set at 0.80 or 0.90. It represents the probability of correctly detecting an effect when it truly exists.
3. Effect Size
It represents the magnitude of the difference or relationship between groups in practical and clinical terms. It is usually estimated based on literature or pilot studies. It is not a fixed value and is specific to each study.
4. Hypothesis Direction
Two-tailed: The standard approach
One-tailed: Used only when the direction of the effect is known in advance with certainty
Step-by-Step Sample Size Calculation Using G*Power
Scenario: Comparing Two Independent Groups (Independent Samples t-test)
Assume that we want to compare the effect of two different treatment methods on patient recovery time.
1. Select the Test Family
Test family: t tests
2. Select the Statistical Test
Statistical test:
Means: Difference between two independent means (two groups)
3. Select Analysis Type
Type of power analysis:
A priori (required sample size)
4. Input Parameters
Tail(s): Two
Effect size (d): Should be calculated using literature or pilot data
α error probability: 0.05
Power (1-β): 0.80 or 0.90
Allocation ratio (N2/N1): 1 (for equal group sizes)
5. Run the Calculation
Click the Calculate button to obtain the required minimum sample size.
Interpreting the Results
The most important output from G*Power is:
Total sample size: Minimum required number of participants
Actual power: The achieved statistical power value
Example Output:
Total sample size: 128
Per group: 64 participants
Drop-out Rate (Attrition)
In clinical and field studies, data loss is common. Therefore:
Adding a 10%–20% drop-out rate is a standard practice
In long-term studies, this rate may be even higher
Example:
If G*Power suggests 100 participants, the study should include approximately 110–120 participants.
Important Note (Jury Perspective)
In thesis defenses, simply stating “I used G*Power” is not enough. Jury members typically evaluate:
How the effect size was determined
Which literature sources were used
Why power was set to 0.80 or 0.90
The justification for the drop-out rate
Conclusion
Accurate sample size calculation directly affects the scientific validity of a study. While tools like G*Power simplify the process, the most critical factor is the scientific justification of statistical parameters used in the analysis.