Bioequivalence and Bioavailability Forum • two different approaches

two different approaches [Nonparametrics]

posted by martin – Austria, 2010-09-16 17:35 (4963 d 13:54 ago) – Posting: # 5919
Views: 20,035

dear d_labes and HS !

I included two additional bootstrap based approaches implemented in R package pairwiseCI where investigation of empirical coverage of bootstrap based CIs is a real stress test (7 hours)

set.seed(150910)

n <- 10

nsim <- 1E4

conf.level <- 0.9

cover <- rep(0,4)

grp <- as.factor(c(rep('A', n), rep('B', n)))

for(i in 1:nsim){

    timeForecast(i, nsim) 

    x <- round(rnorm(2*n, mean=0, sd=1),1)

    res1 <- confint(wilcox_test(x~grp, distribution='exact',

            conf.int=TRUE, conf.level=conf.level))$conf.int

    res2 <- wilcox.exact(x~grp, conf.int=TRUE,

            conf.level=conf.level)$conf.int

    res3 <- Median.diff(x=x[1:n], y=x[(n+1):(2*n)],

            conf.level=conf.level)$conf.int

    res4 <- HD.diff(x=x[1:n], y=x[(n+1):(2*n)],

            conf.level=conf.level)$conf.int

    if(res1[1] <= 0 & res1[2] >= 0){cover[1] <- cover[1] + 1}

    if(res2[1] <= 0 & res2[2] >= 0){cover[2] <- cover[2] + 1}

    if(res3[1] <= 0 & res3[2] >= 0){cover[3] <- cover[3] + 1}

    if(res4[1] <= 0 & res4[2] >= 0){cover[4] <- cover[4] + 1}

  }

-> estimated simulation time <h:mm:ss> (done / all):   0:00:00 / -:--:--

   -:--:-- / 7:02:49  

print(cover/nsim)

[1] 0.9283 0.9372 0.9319 0.9053

the empirical coverage of 0.9053 for the difference of Harrell-Davis estimates for location looks nice.

best regards

martin

PS.: you can interpret the empirical coverage as a proportion based on a sample size of nsim enabling to calculate a CI enabling to assess simulation accuracy

binom.test(x=0.9053*nsim, n=nsim, p=0.9)



        Exact binomial test



data:  0.9053 * nsim and nsim 

number of successes = 9053, number of trials = 10000, p-value = 0.07729

alternative hypothesis: true probability of success is not equal to 0.9 

95 percent confidence interval:

 0.8993926 0.9109701 

sample estimates:

probability of success 

                0.9053

Complete thread:

tmax in case of ties: R vs. R vs. SAS d_labes 2010-09-10 12:05 [Nonparametrics]
- tmax in case of ties: StatXact, Phoenix, and... Helmut 2010-09-10 18:28
  - Ties, no ties, ties, no ties ... d_labes 2010-09-13 13:21
    - simulation martin 2010-09-14 08:33
      - simulants of the world unite d_labes 2010-09-15 10:14
        
        simulants of the world unite martin 2010-09-15 14:53
        
        simulants of the world unite Helmut 2010-09-15 16:17
        
        simul ants with no ties d_labes 2010-09-16 11:19
        
        simul ants with no ties Helmut 2010-09-16 14:01
        
        two different approachesmartin 2010-09-16 15:35
- evaluation of tmax: use of relative effects? martin 2010-09-16 18:04
  - Not positive about that Helmut 2010-09-16 19:52
    - Not positive about that martin 2010-09-16 21:00
      - Stupidity Helmut 2010-09-17 13:10
- tmax in case of ties: R vs. R vs. SAS Jack 2010-09-20 14:03
  - R packages Helmut 2010-09-20 14:20
    - R packages code d_labes 2010-09-27 11:27
    - R packages Jack 2010-09-27 16:14
  - Is Exact exact? d_labes 2010-09-27 09:53