My committee says my sample size is too small. Now what?

By Dr. Rafiq Muhammad, MD, PhD · Updated November 2026

When a committee says your sample is too small, they rarely mean a specific number. They mean you have not justified the number you have. That is fixable, often without recruiting a single extra participant.

What they are actually asking

Behind "too small" sits one of three worries: the study cannot detect the effect you care about (quantitative), the sample cannot support the claims you make (qualitative), or you chose the number for convenience and said nothing about it. Find out which one before you answer. Ask: "Is the concern statistical power, or how far I can generalise?"

Route 1: run an a priori power analysis

If you have not collected data yet, fix four things and solve for the fifth: the test you will run, alpha (usually .05), power (usually .80), and a target effect size. Take the effect size from a prior study or a meta-analysis, or define the smallest effect that would matter in practice. Never pick it because it gives a convenient number. Then add your expected dropout on top.

Lakens (2022) sets out six accepted ways to justify a sample size. Naming the one you used is often enough to end the conversation.

Route 2: you cannot recruit more? Run a sensitivity analysis

If the sample is fixed (a rare population, a closed cohort, data already collected), turn the question around: with this N, alpha and 80% power, what is the smallest effect you could reliably detect? Report that number and say whether effects of that size are plausible in your field. This is honest, and committees accept it far more readily than silence.

Do not compute "observed" or post-hoc power from your own result. It only restates your p-value, and examiners know it.

If even the sensitivity analysis looks weak, change the design rather than the story: a within-subjects design, a strong covariate, or a more reliable outcome measure all raise power at the same N.

Route 3: qualitative study? Argue information power, not numbers

For interviews or focus groups the question is whether your sample holds enough information for your aim. Malterud, Siersma and Guassora (2016) give five factors: how narrow the aim is, how specific the participants are, whether you use established theory, the quality of the dialogue, and your analysis strategy. A narrow aim with highly specific participants needs fewer people. Say which factors apply and point to your saturation or information-power evidence.

A paragraph you can adapt

The target sample of N = [n] was set by an a priori power analysis for [test] (alpha = .05, power = .80), assuming an effect of [d = x] based on [source]. Allowing for [y]% attrition, we will recruit [n2]. A sensitivity analysis shows the achieved sample can detect effects of [d = z] or larger.