How to Calculate Sample Size for Your Study: A Practical Guide

Sample size is one of the few methodology decisions that can make or break a study before data collection begins. Too few participants and your study won't have the statistical power to detect the effect you're looking for, even if the effect is real. Too many and you've wasted time and resources collecting data you didn't need. You may also have exposed more participants to research procedures than necessary. Calculating the right sample size isn't a guess. It's a calculation based on what you're trying to find, how confident you want to be, and what kind of study you're running.

This guide walks through how to calculate sample size for both quantitative and qualitative research. It covers a worked example using G*Power, sample size guidelines by study type, and what to do when you can't get the recommended number. For the broader methodology context, see our research methodology guide. For the foundational concepts of population and sample, see our companion guide on population vs sample in research.

Why Sample Size Matters

Sample size determines whether your study can answer the question you set out to ask. In quantitative research, sample size determines statistical power, which is the probability that your study will detect a real effect when one exists. A study with 80% power has an 80% chance of finding a true effect of the size you specified. It has a 20% chance of missing it. That miss is a Type II error. Studies with low power produce non-significant results that don't actually demonstrate the absence of an effect. They demonstrate that the study wasn't capable of finding it.

In qualitative research, sample size determines whether you've heard enough perspectives to understand the phenomenon you're studying. Too few participants and you risk drawing conclusions from idiosyncratic experiences. Too many and you're collecting interview data that won't add to your understanding while burning time and resources. The qualitative answer to "how many" is "enough to reach saturation," and the practical question is how to plan for that.

Reviewers screen for sample size adequacy in the methodology section. A sample too small to detect the predicted effect is one of the most common reasons for revise-and-resubmit decisions. So is a qualitative sample that hasn't been justified relative to the research question.

Calculating Sample Size for Quantitative Studies

For quantitative research, sample size is calculated through a power analysis. The power analysis takes the inputs you provide and returns a minimum sample size. That number is what you need to detect an effect of the size you specified, at the confidence you specified. The four inputs are non-negotiable: every quantitative power analysis requires them.

The Four Inputs to a Power Analysis

Where Effect Size Estimates Come From

Effect size is the input most graduate researchers struggle with, because it asks you to estimate something you don't know yet. Three sources are commonly used. For what these measures are and how to report them once you have results, see the guide to effect sizes.

• Prior published research. If similar studies have been published, look up their effect sizes and use the average or median as your estimate. This is the strongest approach when good prior research exists.
• Pilot data. If you've run a small pilot study, calculate the effect size from your pilot results. Pilot effect sizes are often unstable due to small samples, so treat them as ballpark estimates rather than precise values.
• Cohen's conventions. When neither prior research nor pilot data is available, use Cohen's small (0.2), medium (0.5), or large (0.8) conventions. Smaller assumed effect sizes require larger samples, so a small effect size assumption produces a more conservative estimate.

A common error is to assume a large effect size to keep the calculated sample size manageable. This produces a study that's underpowered if the actual effect is medium or small. The honest approach is to use the smallest effect size that would be theoretically or practically interesting, and recruit accordingly.

A Worked Example Using G*Power

Suppose you're planning an experiment comparing an intervention group and a control group on a continuous outcome. You expect a medium effect size, you want 80% power, and you'll use a two-tailed alpha of 0.05. Here's how the calculation runs in G*Power, a free power analysis tool widely used in social science and health research.

1. Open G*Power and select the test family. For a two-group comparison on a continuous outcome, choose "t tests" from the test family menu.
2. Select the specific test. Choose "Means: Difference between two independent means (two groups)."
3. Set the analysis type. Select the a priori option, which computes the required sample size from alpha, power, and effect size.
4. Enter the inputs. Effect size d = 0.5 (medium). Alpha = 0.05. Power = 0.80. Allocation ratio N2/N1 = 1 (equal groups).
5. Read the output. G*Power calculates that you need 64 participants per group, or 128 total, to detect a medium effect with 80% power at an alpha of 0.05.
6. Adjust for expected attrition. If you expect 10% dropout, increase your recruitment target by that proportion. 128 divided by 0.90 is 143 participants total.

If you change any input, the required sample size changes. Increasing power to 0.90 raises the requirement to 86 participants per group. Decreasing the assumed effect size to 0.3 raises it to 176 per group. That sensitivity is the point: the power analysis tells you what your specific design requires given your specific assumptions. Document the inputs you used so reviewers can evaluate the choices.

Calculating Sample Size for Qualitative Studies

Qualitative sample size doesn't follow power analysis logic, because qualitative research isn't trying to estimate effects in a population. The goal is to understand a phenomenon in depth, and the question is how many participants that understanding requires. The concept that does this work is saturation.

What Saturation Means and How to Plan for It

Saturation is the point at which additional data collection stops yielding new information. In an interview study, you reach saturation when the next interview produces themes you've already heard, perspectives you've already documented, and explanations you've already coded. Continuing past that point increases your data volume without increasing your understanding.

The challenge is that you can't predict exactly when saturation will arrive before you start collecting data. The honest answer in your methodology section is to specify a planned sample size range based on the literature for similar studies. Then commit to assessing saturation iteratively during collection, and adjust upward if you haven't reached it at the planned endpoint.

Typical Qualitative Sample Sizes by Study Type

More homogeneous samples reach saturation faster. A study of first-year international graduate students at one US university typically saturates around 12 to 20 interviews, because the participants share important characteristics. A study of working adults from diverse backgrounds in multiple cities may take 30 or more interviews, because the sample is more heterogeneous.

Sample Size by Study Type: Quick Reference

The table below summarizes typical sample size ranges across common study types. Treat these as starting points, not as substitutes for a proper power analysis or saturation assessment specific to your study.

What to Do When You Can't Get the Recommended Sample Size

Real-world constraints often prevent researchers from collecting the ideal sample. The right response depends on the study type and on how big the gap is between what you can collect and what would be ideal.

• Reframe the research question. If your sample is too small to detect a small effect, narrow the question to a larger effect your sample can detect. A small study that answers a focused question well beats one that fails to answer a broader question.
• Use a more powerful design. Within-subjects designs, where participants serve as their own controls, typically require fewer participants than between-subjects designs to detect the same effect. Repeated measures designs increase power for the same sample size.
• Use a more sensitive outcome measure. Continuous outcomes generally require smaller samples than dichotomous outcomes for equivalent statistical power. If your hypothesis allows it, prefer continuous measurement.
• Acknowledge the limitation transparently. If you can't recruit the recommended sample and can't change the design, name the limitation in your methodology and discussion sections. Reviewers trust studies that acknowledge limitations honestly.
• Reframe as a pilot study. If your sample is too small to support definitive conclusions, frame the study explicitly as preliminary or exploratory. Pilot studies have legitimate scientific value when reported as such, including informing power analyses for future larger studies.

How to Justify Sample Size in Your Methodology Section

Reviewers expect to see sample size justified, not just stated. The methodology section needs to explain why you chose the number you did. Strong justifications follow a predictable structure.

1. State the sample size. Specify the number of participants enrolled, the number who completed the study, and any attrition.
2. Cite the calculation method. For quantitative studies, name the power analysis tool and cite the source. For qualitative studies, cite the saturation literature for similar studies.
3. Specify the inputs. For quantitative studies, report the assumed effect size, statistical power, alpha level, and statistical test. For qualitative studies, describe how saturation was assessed.
4. Justify the inputs. Cite prior research, pilot data, or convention for the assumed effect size. Cite established standards for power and alpha, typically 0.80 and 0.05.
5. Address attrition and missing data. If your final analytic sample is smaller than enrollment, explain why and discuss whether the loss was random or systematic.
6. Discuss the limits. If your sample is smaller than ideal or doesn't perfectly fit the planned analysis, address the implications in the limitations section.

Common Sample Size Mistakes

The same problems appear in graduate research over and over. Knowing them in advance saves a round of revisions.

• Assuming a large effect size to keep the sample manageable. This produces a study that's underpowered if the actual effect is medium or small. Use the smallest effect size that would be theoretically interesting, not the largest you can defend.
• Running the power analysis after data collection. Post-hoc power based on observed effect sizes tells you nothing about whether the original study was adequately powered. The power analysis must come first.
• Treating qualitative sample size as a number to hit rather than a process to follow. Qualitative sample size depends on saturation, which depends on the data. Specify a planned range and assess saturation iteratively.
• Confusing descriptive and inferential sample size needs. A sample of 30 may be plenty to describe a specific group. It's typically too small to support generalization to a larger population.
• Forgetting attrition. Recruit enough participants to end up with the analytic sample your power analysis requires. If you expect 10% dropout, recruit 10% extra at minimum.
• Ignoring clustering or nesting in the design. If participants are nested within groups, such as students within schools, the effective sample size for some analyses is smaller than the participant count. Multilevel power analysis is required.

Frequently Asked Questions

Professional Editing for Your Research Manuscript

Sample size justification is one of the methodology elements reviewers screen for first. A vague or incomplete discussion can trigger a desk rejection before reviewers evaluate the rest of the study. Be clear about your power analysis inputs, the source of your effect size estimates, and the limits of your final sample. That clarity is one of the strongest signals that the rest of the methodology is rigorous.

Editor World provides dissertation editing and academic editing services for researchers preparing theses, dissertations, and journal submissions. Every editor is a native English speaker from the United States, the United Kingdom, or Canada, with an advanced degree in their field. Every document is reviewed by a real person, never by AI. You can choose your own editor from the Editor World roster. You can also request a free sample edit of your first 300 words before committing. Pricing is transparent through an instant price calculator that shows your exact cost upfront.

A certificate of editing confirming human-only native English editing is available as an optional add-on for journal submissions where AI use must be disclosed. For more on research methodology and design, see our research methodology guide and population vs sample in research. Also useful are quantitative vs qualitative research and the research bias guide.

Page last reviewed: September 2026. Content reviewed by Editor World editorial staff. Editor World, founded in 2010 by Patti Fisher, PhD, graduate of The Ohio State University, provides professional editing and proofreading services for academic researchers, doctoral candidates, faculty, business professionals, and authors worldwide. 100% human editing, no AI at any stage. BBB A+ accredited since 2010 with 5.0/5 Google Reviews and 5.0/5 Facebook Reviews. 16 years in business with 140 million+ words edited for over 8,000 clients in 65+ countries. Multiple Gold and Bronze Stevie Award winner. Native English editors from the United States, the United Kingdom, and Canada. Less than 5% of applicants are accepted to the editor panel. Recommended by the Boston University Economics Department, University of San Diego, University of Michigan, UCLA, University of Missouri, and more.