Skip to content
Omnikalc

Dropout sample size calculator

Enter the number of completers you need per group, the expected dropout rate, the effect size and the number of groups to get the number to enroll, the common mistake of simply adding the percentage, and the power you would lose without inflating.

Updated on
Your inputs

Value out of range

%

Value out of range

Value out of range

Value out of range

Results

Participants to enroll per group

118

Total to enroll
236
Extra participants per group
18
Naive count (adding the percentage)
115
The naive count leaves you short of completers
Planned power with the completers
80.7 %
Power if you enroll only the planned number
74.1 %
Below the usual 80 %
Loss of power
6.6 points

The dropout sample size calculator gives the number to enroll to keep enough completers, shows why simply adding the percentage falls short, and estimates the power lost if you do not inflate the sample.

How the calculation works

  • Enter the completers needed per group and the dropout rate.
  • Enter the effect size and the number of groups.
  • The calculator gives the number to enroll, the total, the naive count and the power with and without inflation.

Example

You need 100 completers per group, expect 15 % dropout and an effect of d = 0.4. You must enroll 100 / 0.85 = 118 per group, or 236 in total. Adding 15 % gives only 115. With 100 completers, the power is 80.7 %, but if you enroll only 100 and lose 15 %, the 85 completers give a power of only 74.1 %, a loss of 6.6 points.

Why it matters

The calculator helps clinical researchers who plan trials and cohort studies, students who write protocols, and funders who check whether a sample size is realistic.

Good to know

This inflation covers loss of participants who do not provide the outcome, whatever the reason. It does not remove the bias that arises if the people who drop out differ from those who stay, so plan efforts to limit dropout and analyze it. The power formula is a normal approximation for a comparison of two means, and the true power for a t-test is slightly lower for small samples. For three or more groups or other designs, apply the same inflation to the number of completers computed by a suitable power calculation. Round up to whole participants in every group, and consider a margin for the unforeseen.

The statistical power calculator links effect, sample and power, the sample size calculator sizes surveys, the Cohen’s d calculator gives effect sizes and the non-inferiority sample size calculator plans non-inferiority trials.

Frequently asked questions

How do I inflate a sample for dropout?

Divide the number of completers needed by (1 − dropout rate): enroll n / (1 − d). For 100 completers and 15 % dropout, that is 100 / 0.85 = 118, not 115. Adding the percentage ignores that the dropout applies to the enrolled number, not to the final number.

How is the power computed?

For two groups with n completers each, power ≈ Φ(d × √(n / 2) − 1.96) at a two-sided 5 % level, using a normal approximation. Fewer completers lower the term in the bracket and so the power.

How do I estimate the dropout rate?

Look at similar trials with the same follow-up length and population, or at a pilot. Dropout is often larger than hoped, and may differ between groups, which can bias the results as well as lower power.

Was this calculator helpful?

Thanks for your feedback!

Results are estimates provided for information purposes. They do not replace professional advice.