# Calculating sample size

#### Goocha

##### New Member
Hello,
I am going to conduct a research and would like to estimate the sample size that I will need.
Here is the information that I have, please let me know if you need some more information.

We know from previous research that the probability of phenomena in one group (control group) is 1.1% and in another independent group (research group) is 19.4%.
I would like to check, what are the differences between those 2 groups in several demographics characteristic by using t-test and non parametric test. Alpha of 0.05%

What is the sample size that I should take?

#### Karabiner

##### TS Contributor
I would like to check, what are the differences between those 2 groups in several demographics characteristic by using t-test and non parametric test. Alpha of 0.05%
should be in order to have enough power (how much do
you want?) when testing the frequency of the phenomenon?
Or, how large it should be in order to test for differences
in sociodemographic characteristics? If it is the first, then
you could use the g*power freeware or do a websearch
for an online-calculator for 2x2 tables with Chi² test.

With kind regards

K.

#### Goocha

##### New Member
Why question is how large my sample should be in order to have enough power (mediud size) when testing the frequency of the phenomenon?

#### Karabiner

##### TS Contributor
So you have 2 groups and a binary dependent variable (phenomenon
obeserved yes/no)? I don't know what you mean by "medium power"
exactely, but as I said before, if you know the effect size (1.1% versus
19.4%), the desired power, alpha = 5%, and the test (Chi²), then
you could use the g*power freeware http://www.gpower.hhu.de/, or
have a look at http://powerandsamplesize.com/

With kind regards

K.