[R] sample size 2-sample proportion test

From: Berta <ibanez_at_bioef.org>
Date: Tue, 17 Apr 2007 14:10:47 +0200

Hi R-users,
I want to calculate the sample size needed to carry out a 2-sample proprotion test (with alfa=0.05, beta=0.8)

```1.- probability of success in subpopulation A: 0.8
2.- probability of success in subpopulation B: 0.05
3.- percentage of population in subpopulation A = 5%,
4.- percentage of population in subpopulation B= 95% (the rest, A and B
```
complementary)
(Note that it is dificult to find people from subpopulation A)

Using library(Hmisc) (or others if exist?), I would like to know which one of these three options (if any) should I use:

# FIRST:

bsamsize(p1=0.8, p2=0.05, fraction=0.05, alpha=.05, power=.8)
# n1 n2
# 1.46 27.82

# SECOND:

samplesize.bin(alpha=0.05, beta=0.8, pit=0.8, pic=0.05, rho=0.05)
# n= 42

# THIRD: To calculate sample size assuming fraction=0.5 and then multiplying
by 0.05:
bsamsize(p1=0.8, p2=0.05, fraction=0.5, alpha=.05, power=.8)
# n1 n2

# 5.47 5.47
# so that n=6*100/5=120

Taking into account the difficulty of obtaining the samples, which method would be recomended? Or in other words, how many people should I find, around 30-40 or 120?