bootstrap for confidence intervals of the mean

d = c(0L, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, 0L, 0L, 7375L,
NA, NA, 17092L, 0L, 0L, 32390L, 2326L, 22672L, 13550L, 18285L)

boot.out <-boot(d, mean, R=1000, sim="permutation")

Error in mean.default(data, original, ...) :
  'trim' must be numeric of length one

I know that I am missing something but I can't figure it out.

You aren't reading the documentation closely enough. ?boot informs us
that for all sim other than "parametric", 'statistic' must be a function
that takes two arguments. The second argument to mean.default is trim,
and boot is passing to mean a vector of indices as argument 'trim' which
it is not expecting and quite rightly throws a wobbly.

Write a wrapper to mean, that accepts two arguments, one the data vector
and one the permuted indices, then use these to form a call to mean ---
here mean.fun does this (note we turn on removing NA's by default
otherwise it wouldn't work)

mean.fun <- function(dat, idx) mean(dat[idx], na.rm = TRUE)
boot.out <- boot(d, mean.fun, R=1000, sim="ordinary")
boot.out

ORDINARY NONPARAMETRIC BOOTSTRAP


Call:
boot(data = d, statistic = mean.fun, R = 1000, sim = "ordinary")


Bootstrap Statistics :
    original    bias    std. error
t1* 9474.167 -3.373422    3110.968

There doesn't seem to be much point in permuting the data here, the mean
will be the same, regardless of what permutation you take. The above
does an ordinary bootstrap instead.

HTH

G

thanks

stephen

%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%
 Dr. Gavin Simpson             [t] +44 (0)20 7679 0522
 ECRC, UCL Geography,          [f] +44 (0)20 7679 0565
 Pearson Building,             [e] gavin.simpsonATNOSPAMucl.ac.uk
 Gower Street, London          [w] http://www.ucl.ac.uk/~ucfagls/
 UK. WC1E 6BT.                 [w] http://www.freshwaters.org.uk
%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%