The prop_power() function from the catfun package can be used for conducting power analyses for comparisons of two proportions in R.
The prop_power() function can be used to calculate power for equal group sample sizes.
prop_power(n = 220, p1 = 0.35, p2 = 0.2, alpha = 0.05)##
## Two-sample comparison of proportions power calculation
## ---------------------------------------------------------------------
##
## Total sample size: 220
## N1: 110
## N2: 110
## Proportion 1: 0.35
## Proportion 2: 0.2
## Power: 0.705077
## Significance level: 0.05The prop_power() function can also be used to calculate sample size given power. Instead of specifying the number of observations per group or the total sample size, we specify the desired power.
prop_power(p1 = 0.35, p2 = 0.2, power = 0.80, alpha = 0.05)##
## Two-sample comparison of proportions power calculation
## ---------------------------------------------------------------------
##
## Total sample size: 275.8295
## N1: 137.9148
## N2: 137.9148
## Proportion 1: 0.35
## Proportion 2: 0.2
## Power: 0.8
## Significance level: 0.05The prop_power() function can be used to calculate unbalanced sample sizes. The fraction argument is set to 2/3 indicating that 2/3 of the total sample will be members of group one.
prop_power(p1 = 0.35, p2 = 0.2, fraction = 2/3, alpha = 0.05, power = 0.85)##
## Two-sample comparison of proportions power calculation
## ---------------------------------------------------------------------
##
## Total sample size: 359.6202
## N1: 239.7468
## N2: 119.8734
## Proportion 1: 0.35
## Proportion 2: 0.2
## Power: 0.85
## Significance level: 0.05Power analyses in SAS are conducting using the power procedure.
title1 'Computing Power Given Sample Size';
proc power;
twosamplefreq test = pchi
alpha = 0.05
groupproportions = (0.35 0.20)
nullproportiondiff = 0
groupweights = (1 1)
ntotal = 220
power = .;
run;## Computing Power Given Sample Size
##
## The POWER Procedure
## Pearson Chi-square Test for Proportion Difference
##
## Fixed Scenario Elements
##
## Distribution Asymptotic normal
## Method Normal approximation
## Null Proportion Difference 0
## Alpha 0.05
## Group 1 Proportion 0.35
## Group 2 Proportion 0.2
## Group 1 Weight 1
## Group 2 Weight 1
## Total Sample Size 220
## Number of Sides 2
##
##
## Computed Power
##
## Power
##
## 0.705title1 'Computing Sample Size Given Power';
proc power;
twosamplefreq test = pchi
alpha = 0.05
groupproportions = (0.35 0.20)
nullproportiondiff = 0
npergroup = .
power = 0.80;
run;## Computing Sample Size Given Power
##
## The POWER Procedure
## Pearson Chi-square Test for Proportion Difference
##
## Fixed Scenario Elements
##
## Distribution Asymptotic normal
## Method Normal approximation
## Null Proportion Difference 0
## Alpha 0.05
## Group 1 Proportion 0.35
## Group 2 Proportion 0.2
## Nominal Power 0.8
## Number of Sides 2
##
##
## Computed N per Group
##
## Actual N per
## Power Group
##
## 0.800 138title1 'Computing Sample Size Given Power (Unbalanced Groups)';
proc power;
twosamplefreq test = pchi
alpha = 0.05
groupproportions = (0.35 0.20)
nullproportiondiff = 0
groupweights = (2 1)
ntotal = .
power = 0.85;
run;## Computing Sample Size Given Power (Unbalanced Groups)
##
## The POWER Procedure
## Pearson Chi-square Test for Proportion Difference
##
## Fixed Scenario Elements
##
## Distribution Asymptotic normal
## Method Normal approximation
## Null Proportion Difference 0
## Alpha 0.05
## Group 1 Proportion 0.35
## Group 2 Proportion 0.2
## Group 1 Weight 2
## Group 2 Weight 1
## Nominal Power 0.85
## Number of Sides 2
##
##
## Computed N Total
##
## Actual N
## Power Total
##
## 0.850 360