Compute bias-corrected, accelerated confidence intervals for non-parametric bootstrapped parameter estimates.

Usage

bca_jacknife(boot.est, OG.dat, est.fxn, alpha, multcomp = TRUE)

Arguments

boot.est: A data frame with n x p rows (n= # of bootstrap replicates, p= # of model terms or parameters and two columns: one for the parameter name (term), and one for the parameter estimate (estimate))
OG.dat: The orignal data set
est.fxn: The summary function to be performed on the data. See est_accelerate for a more detailed description of what this function needs to return.
alpha: The desired 2-tail alpha level
multcomp: If true, the resulting table will have adjusted CI's with the family-wide error rate at the specified alpha level. If FALSE, the error rate is applied to each parameter. Correction using the Sidak method is instituted by default. This is only appropriate for when you are looking for a table of regression coefficients and their bootstrapped CI's.

Value

A data frame containing the parameter name (term), theta.hat, the estimate of the parameter from the orignal data, and the lower and upper BCa confidence intervals

Examples


# data.frame(X1=seq(1,10,1),
#            X2=rnorm(10,0,1))%>%
#   mutate(Y=0.2*X1+X2)%>%
#   select(-X2)->tmp
#
# Create bootstrapped data sets
#
# purrr::map(1:10,~sample(1:nrow(tmp),size=nrow(tmp),replace=TRUE)%>%
#               tmp[.,])->boot.dat
#
#
# my.fxn1<-function(X){
#   lm(Y~X1,data=X)%>%
#     broom::tidy(.)%>%
#     select(term,estimate)
# }
#
#
# #this function above could look different every time depending on bootstrap approach
# # but should return a 2-column dataframe with columns term- the parameter name,
# # and estimate. You do not need to use a function here, I just did for convenience.
# # You just need to get the bootstrapped parameter estimates in the correct format.
#
#
# est_boot<-function(est.fxn,boot.list){
#
#   map(boot.list,~est.fxn(.x))%>%
#     bind_rows()%>%
#     rename(theta.boot=estimate)->out
#
#   return(out)
#
# }
#
#
# est_boot(my.fxn1,boot.dat)->boot.out
#
#
# bca_jacknife(boot.out,tmp,my.fxn1,alpha=0.05,multcomp = FALSE)