Binomial test and chi^2-test for proportions

Install required packages

DescTools

wants <- c("DescTools")
has   <- wants %in% rownames(installed.packages())
if(any(!has)) install.packages(wants[!has])

Binomial test

One-sided

DV   <- factor(c("+", "+", "-", "+", "-", "+", "+"), levels=c("+", "-"))
N    <- length(DV)
(tab <- table(DV))

DV
+ - 
5 2 

pH0 <- 0.25
binom.test(tab, p=pH0, alternative="greater", conf.level=0.95)

    Exact binomial test

data:  tab
number of successes = 5, number of trials = 7, p-value = 0.01288
alternative hypothesis: true probability of success is greater than 0.25
95 percent confidence interval:
 0.3412614 1.0000000
sample estimates:
probability of success 
             0.7142857 

Two-sided

N    <- 20
hits <- 10
binom.test(hits, N, p=pH0, alternative="two.sided")

    Exact binomial test

data:  hits and N
number of successes = 10, number of trials = 20, p-value = 0.01704
alternative hypothesis: true probability of success is not equal to 0.25
95 percent confidence interval:
 0.2719578 0.7280422
sample estimates:
probability of success 
                   0.5 

sum(dbinom(hits:N, N, p=pH0)) + sum(dbinom(0, N, p=pH0))

[1] 0.01703563

Confidence intervals

library(DescTools)
BinomCI(tab[1], sum(tab), method=c("wilson", "jeffreys", "midp"))

               est    lwr.ci    upr.ci
wilson   0.7142857 0.3589345 0.9177811
jeffreys 0.7142857 0.3523383 0.9352717
midp     0.7142857 0.3302059 0.9490209

\(\chi^{2}\)-test for proportions

total <- c(4000, 5000, 3000)
hits  <- c( 585,  610,  539)
prop.test(hits, total)

    3-sample test for equality of proportions without continuity
    correction

data:  hits out of total
X-squared = 50.587, df = 2, p-value = 1.035e-11
alternative hypothesis: two.sided
sample estimates:
   prop 1    prop 2    prop 3 
0.1462500 0.1220000 0.1796667 

Detach (automatically) loaded packages (if possible)

try(detach(package:DescTools))

Get the article source from GitHub

R markdown - markdown - R code - all posts