R语言课后练习求大神解答！

楼主的妹子是学统计的，以下代码是她们老师今天在课堂上讲的，求power的function。

老师留的作业是用power反推somple size，妹子已经哭得做不出来了，来这求大神们帮忙给点思路！！！！万分感谢！！！

###Example 1

#Let P(D|X=0) = 0.17 & P(D|X=1) = 0.34

#Let proposed n1. = 50 and let n2. = 50

n1. <- 50

n2. <- 50

reps <- 10000 #how many times we will resample and conduct hypothesis test

p0 <- 0.17

p1 <- 0.34

alpha = 0.05

pvals <- rep(NA, reps) #vector that will store the p-values

seed <- 1234 #ensure reproducible results

#now conducting the simulation

chisq.power <- function(n1., n2., reps, p0, p1, alpha, pvals, seed){

set.seed(seed)

for (i in 1:reps){
        x0 <- rbinom(n=n1., size=1, prob=p0) # simulating n1. bernouli rv's w/ Pr(D) = p0
        x1 <- rbinom(n=n2., size=1, prob=p1) # simulating n2. bernouli rv's w/ Pr(D) = p1
        cross.tab <- matrix(c(sum(x0==1), sum(x0==0), sum(x1==1), sum(x1==0)), ncol=2, byrow=T)
        pvals[i] <- chisq.test(cross.tab, correct=FALSE)$p.value
}
return(pvals)
}

results <- chisq.power(n1., n2., reps, p0, p1, alpha, pvals, seed)
Power <- length(which(results<alpha))/reps
Power


接上代码

不懂帮顶

wocaishiliuking 发表于 2017-2-15 14:21
不懂帮顶

谢谢关注

老师的原问题是：How might you revise this function to calculate the sample size needed for 80% power assuming