explain inference with boot strap and not null-hypothesis testing

[floswald]

• https://cran.r-project.org/web/packages/dabestr/
• https://youtu.be/5Dnw46eC-0o
• http://pages.stat.wisc.edu/~larget/stat302/chap3.pdf

[floswald]

• generate a bootstrapped sample by collecting a slope estimate from each student.
• each student does

library(dplyr)
new_cars = sample_n(cars, size = nrow(cars), replace = TRUE)
b1 = coef(lm(speed ~ dist, new_cars))[2]

• collect values somehow. google sheets?
• make each student do it 5 times
• then draw histogram of all collected b1's

boostrap

library(bootstrap)
fc = function(d,i){
d2 = d[i, ]
return(coef(lm(speed ~ dist,data = d2))[2])
}
bootcars <- boot(cars, fc, R=500)
plot(bootcars)

relate to testing against zero