STATISTICAL SOFTWARE
G*Power Sample Size and Power Analysis
Justified Inputs, Documented Settings
Our G*Power sample size and power analysis matches the test family to the model you plan to fit, takes the effect size from a defensible target and records alpha, power, tails and attrition, so that a grant reviewer or ethics board can reproduce every figure.
No Commitment Required
Strictly Confidential
In-House PhD-Level Statisticians
Fixed Project Fee
Same-Day Response
What Is G*Power Used For in Research?
G*Power is a freely available program for statistical power analysis, developed by researchers at Heinrich Heine University Düsseldorf. It computes the sample size required, the power achieved or the effect size a study can detect for common tests, including t-tests, ANOVA, regression, correlation and chi-square tests. Researchers use it because grant applications, ethics submissions and journals expect a documented sample size justification, and because each calculation names its test family, effect size and error rates. OnlineSPSS, a specialist statistical consultancy for academic, clinical and professional research, provides G*Power sample size and power analysis in which an in-house PhD-level statistician sets up the calculation for your planned model and documents every input.
WHEN TO ENGAGE
Four Occasions When Researchers Bring Us a Power Question
A G*Power analysis service is requested before data collection, during a proposal or after a reviewer questions the sample, and each situation needs a different calculation.
Sample Size for a Grant or Protocol A grant, protocol or ethics submission needs a sample size justification tied to the primary analysis, the smallest effect worth detecting and an allowance for participants who drop out. If the design itself is still open, the calculation belongs inside Research Planning & Study Design.
Test Family Unclear The planned model does not map cleanly onto a test family, because of a repeated design, covariates or a clustered sample, and the calculation must match the analysis, not the nearest menu entry.
Reviewer Questions the Sample A reviewer asks where the effect size came from or whether the study was adequately powered, and the justification has to be rebuilt from inputs that can be defended.
Sample Already Fixed Recruitment or resources have set the sample, and a sensitivity analysis states the effect the study could detect, in place of a power figure computed from the observed result.
CALCULATIONS IN G*POWER
Power Calculations Set Up for the Test You Plan to Run
Each calculation follows the statistical test planned for the primary outcome, and its card links to the method page where that test and its assumptions are explained.
Independent Samples t-Test
Two-group designs use the t-test family with a standardized difference taken from prior evidence or from a difference that matters, and the allocation ratio and the tails are stated.
One-Way ANOVA
Several groups use the F-test family with Cohen's f, and the sample reflects the planned pairwise comparisons, because the omnibus test does not by itself provide power for each pairwise contrast.
Repeated Measures ANOVA
Within-subject designs need the number of measurements, the correlation among them and a sphericity correction, all of which change the sample for the same effect size.
Multiple Linear Regression
An overall model uses f-squared and the number of predictors, while a single coefficient or an increase in R-squared has its own variant, chosen to match the hypothesis.
Logistic Regression
A binary outcome needs the baseline probability, the odds ratio worth detecting and the distribution of the predictor, and the expected number of events is checked against the planned number of predictors.
Chi-Square Test
Tests of proportions and contingency tables use an effect size computed from the expected cell proportions, with exact tests considered when expected counts are small.
COMMON POWER ANALYSIS ISSUES
Where Power Calculations Go Wrong Before the Number
A power calculation returns a number whatever is entered, so each input below is checked against the planned design before a sample size is written into a proposal.
Test Family Matches the Model A calculation run in the wrong family returns a plausible number for a different analysis. The family and test type are chosen to match the model that will actually be fitted.
Conventional Benchmarks as Defaults Small, medium and large labels are conventions, not evidence. The effect size comes from a clinically or substantively meaningful difference, or from prior studies adjusted for their likely optimism.
Effect Size Metric Mismatch Published effects arrive as d, partial eta-squared or an odds ratio, and the G*Power effect size field expects the metric of the chosen test, so each conversion is made explicitly and recorded.
Post Hoc Power From the Observed Effect Power computed from the observed effect only restates the P value: a nonsignificant result always returns an observed power below about 50 percent. A reviewer's request for achieved power is answered with the confidence interval and the effect the sample could detect.
Attrition and Non-Response The calculated number is the sample that will be analyzed, so the enrollment target is that number inflated for the loss expected, with the assumption stated.
Designs Outside the Test Families Clustered, multilevel and latent variable designs have no direct equivalent among the test families. A design effect, a simulation or specialist software is used and its assumptions are documented.
DELIVERABLES
What You Receive From a Power Analysis
The documents are written for a proposal, protocol or manuscript, agreed in the scope and provided where applicable, so that a reviewer can rerun the G*Power sample size calculation from your description.
Power Analysis Report A report stating the test family and test, alpha, power, tails, groups or predictors, allocation ratio and the exact G*Power setup behind each figure.
Effect Size Justification The source of each effect size, the conversion applied and the reason it is a defensible target, with the uncertainty around the prior evidence noted.
Sample Size Table The analyzable sample and the enrollment target after attrition across a range of plausible effect sizes and power levels, so the choice is visible.
Sensitivity Analysis For a sample that is fixed, the smallest effect the study can detect at the stated alpha and power, stated so that you can report it among your own limitations.
Methods Paragraph A sample size justification in the wording a protocol, grant or manuscript expects, consistent with the analysis planned for the primary outcome.
Statistical Interpretation The assumptions behind each calculation are explained, together with what the planned test can and cannot show at the sample proposed.
OTHER PLATFORMS
When Another Platform May Be More Appropriate
G*Power covers many standard tests, but a G*Power power calculation follows the design, and some studies need power estimated elsewhere.
ILLUSTRATIVE EXAMPLE
How a Power Analysis May Look in Practice
An illustration of how a first calculation is corrected before it goes into a grant application.
ILLUSTRATIVE EXAMPLE · MOTOR LEARNING LABORATORY RESEARCH · UNITED STATES
A university laboratory planned an experiment comparing two practice schedules on skill retention, with each participant tested at three time points, and needed a sample size justification for a grant application.
The first calculation used a one-way ANOVA between groups with a conventional medium effect size, which was simple to run but did not match the planned analysis. The primary hypothesis was the interaction of schedule and time, so the calculation was redone with the repeated measures, within-between interaction test of the F-test family, entering the number of measurements and the expected correlation between repeated tests. The effect size was replaced by the smallest difference in retention the laboratory considered meaningful, supported by pilot data and a published study, with the optimism of the pilot acknowledged. An allowance for participants who would miss a session was added to the analyzable sample.
Outcome: The laboratory received a power analysis report with the exact setup, an effect size justification, a table of required enrollment across plausible effect sizes and a methods paragraph ready for the application.
This example does not describe a real client or project. We never share details of actual client work, in line with the confidentiality we promise every researcher. This illustrative example is included to show the type, scope, and standard of work involved in this kind of analysis.
RESEARCH DISCIPLINES
Research Fields Where G*Power Is Common
G*Power is widely used in psychology and the behavioral sciences, and every field that writes a protocol or applies for funding needs the same documented justification of its sample.
HOW WE ASSESS YOUR PROJECT
We Review Your Actual Research
Your assessment is based on a review of your actual research materials and analytical requirements, not a generic project category.
You Submit Your Project Details
Share whatever is available at the current stage, such as the research aim, protocol or proposal, dataset, variable list, manuscript or statistical requirements.
We Review Your Research and Define the Scope
Your statistically trained Research Coordinator reviews the materials to understand the study design, research questions, data structure, analytical requirements and expected deliverables. If clarification is needed, additional documents or information can be requested before the assessment is finalized.
Your Analysis Is Completed and Delivered
Once the scope and fixed project fee are agreed, the in-house statistician completes the work and the Research Coordinator remains the communication point through delivery, with the analysis, deliverables and timeline as agreed.
RESEARCHER VOICE
What Researchers Say
Feedback from a principal investigator on choosing the correct test family and effect-size inputs for the planned design, and on documentation the team could rerun.
“Keeping the analysis in G*Power mattered because the rest of the project team needed to rerun it later. They dealt with choosing the correct test family and effect-size inputs for the planned design, documented the analytical steps clearly, and gave us outputs that were much easier to audit and report.”
Statistical Support Within Your Research OnlineSPSS provides the agreed methodological, analytical and statistical reporting support, including the power analysis documents described in the scope. The researcher or research team retains the scientific question, substantive interpretation, conclusions and overall responsibility for the research.
FREQUENTLY ASKED QUESTIONS
Questions About G*Power and Sample Size Planning
Short answers to what researchers ask before a power analysis, from the calculations we set up and the choice of effect size to post hoc power, repeated measures inputs and confidentiality.
Which calculations do you set up in G*Power?
We set up sample size and power calculations for the tests researchers most often plan: t-tests, ANOVA including repeated measures, regression, correlation, logistic regression and chi-square tests. The test family follows the model you plan to fit, and it is confirmed during the assessment.
Where should the effect size come from?
From the smallest effect that would matter clinically or substantively, or from prior studies and pilot data treated with caution, since published and pilot effects tend to be optimistic. A conventional medium value is a last resort and is labeled as one. The source and any conversion are written into the report.
Can you calculate power after my study is finished?
Not from the observed effect, because that figure carries no information beyond the P value. We offer a sensitivity analysis instead: the smallest effect the achieved sample could detect at the stated alpha and power, which is the form reviewers can use.
Why does G*Power give such a small sample for my repeated measures design?
Usually because a partial eta-squared from SPSS was entered under G*Power’s default effect size setting. That value already reflects the correlation between repeated measurements, while the default setting asks for the correlation separately, so the correlation is counted twice and the sample comes out too small. The effect size option is set to match the source of the estimate, and the correlation is taken from pilot or prior data, not left at its default.
How do you protect my research data and confidentiality?
Your protocol, pilot data and planning documents are handled confidentially, and project information is not published. An NDA is available where required and can be arranged before sensitive materials are shared. Read more about confidentiality and data security.
What happens when I request an assessment, and how long does it take?
You share your research question, design and relevant materials, such as the protocol or analysis plan, and a statistically trained Research Coordinator reviews the request and responds the same day. The analysis, deliverables, timeline and one fixed project fee are then confirmed before any work begins, so the timeline follows from the agreed scope.
Still have a question? Find more answers in our full FAQ.
NEXT STEP
Discuss Your
Power Analysis With Us
Share your research question, the design and primary outcome you plan, and any pilot data or prior studies. We check the planned test and confirm the G*Power setup before any work begins.
No Commitment Required
Strictly Confidential
Fixed Project Fee
In-House PhD-Level Statisticians
Same-Day Response
Serving Researchers Since 2014