[banner]

An R Companion for the Handbook of Biological Statistics

Salvatore S. Mangiafico

 

Advertisement

Power Analysis

 

Advertisement

Packages used in this chapter

The following commands will install these packages if they are not already installed:


if(!require(pwr)){install.packages("pwr")}

 

Introduction

Parameters

How it works

See the Handbook for information on these topics.

 

Examples

Power analysis for binomial test

 

### --------------------------------------------------------------
### Power analysis, binomial test, pea color, p. 43
### --------------------------------------------------------------

P0 = 0.75
P1 = 0.78
H  = ES.h(P0,P1)               # This calculates effect size

library(pwr)

pwr.p.test(
       h=H,
       n=NULL,                  # NULL tells the function to
       sig.level=0.05,          #     calculate this
       power=0.90,              # 1 minus Type II probability
       alternative="two.sided")

  

n = 2096.953                 # Somewhat different than in Handbook

 

#     #     #

 

 

Power analysis for unpaired t-test

 

### --------------------------------------------------------------
### Power analysis, t-test, student height, pp. 43
44
### --------------------------------------------------------------

M1  = 66.6                      # Mean for sample 1
M2  = 64.6                      # Mean for sample 2
S1  =  4.8                      # Std dev for sample 1
S2  =  3.6                      # Std dev for sample 2

Cohen.d = (M1 - M2)/sqrt(((S1^2) + (S2^2))/2) 
                                         
library(pwr)
                                  
pwr.t.test(
       n = NULL,                  # Observations in _each_ group
       d = Cohen.d,           
       sig.level = 0.05,          # Type I probability
       power = 0.80,              # 1 minus Type II probability
       type = "two.sample",       # Change for one- or two-sample
       alternative = "two.sided"
       )

 

Two-sample t test power calculation

 

n = 71.61288

 

NOTE: n is number in *each* group 71.61288

 

#     #     #

 

 

How to do power analyses

Methods are shown in the previous examples.