# Sample size calculation

#### Lauraneedshelp

##### New Member
Hi there,

I would like to calculate the sample size for a randomized controlled trial.

Based on a systematic review, the effect of the primary outcome (insomnia severity) is expected to be Hedges's g= 1.03.
The power should be 0.80, alpha is 0.05 (two-tailed).
I am expecting a drop-out rate of 25%.

I found this formula:

If I want to fill in this formula I am missing the standard deviation.
Could someone please advise me on what is the best way to perform this calculation?
I do have Stata 16 aswell where I see options to do this analysis. However, I am not sure which exact method to use.

Best regards,
Laura

#### Attachments

• 146.4 KB Views: 0

#### katxt

##### Active Member
You need to provide an estimate of the SD of the natural variation yourself. Maybe do a preliminary experiment, or look at reported results from a similar investigation.

#### Karabiner

##### TS Contributor
[qute]I would like to calculate the sample size for a randomized controlled trial.[/quote]
Why you? For ethical reasons, perhaps someone with sufficient statistical experience
should carry out the sample size calculations?

If I want to fill in this formula I am missing the standard deviation.
The means and standard deviations are needed to include an expected effect size.

Since you already have an expected effect size (Hedge's g), you just need a formula
(or a software, or online calculator) where you put in g, instead of means and SDs.

With kind regards

Karabiner

#### Lauraneedshelp

##### New Member
Thank you everyone for your suggestions. I found a way to do this calculation in Stata.
Best regards, Laura

#### Lauraneedshelp

##### New Member
[qute]I would like to calculate the sample size for a randomized controlled trial.
Why you? For ethical reasons, perhaps someone with sufficient statistical experience
should carry out the sample size calculations?

The means and standard deviations are needed to include an expected effect size.

Since you already have an expected effect size (Hedge's g), you just need a formula
(or a software, or online calculator) where you put in g, instead of means and SDs.

With kind regards

Karabiner[/QUOTE]

Thank you. This calculation was part of an assignment for a research proposal which I needed to get accepted to a job.
I am aware that I am not an expert in sample size calculations. However, I needed to show that I was able to figure it out. The 'real' sample size calculation will be done together with several people with expertise in this field.
I found an online calculator to do so: https://www.danielsoper.com/statcalc/calculator.aspx?id=47
However, I was hoping to find a more elegant solution, which was then possible in Stata.

Best regards, Laura