[R] Bootstrapping the colMeans statistic

[kim.elmore]

I have a fair bit of experience with S-Plus and have been asked to port 
some of my S-Plus bootstrapping functions to R, to which I am relatively 
new, Needless to say, I've run into some problems. In particular, I need 
to perform bootstrap resampling of the colMeans function using a moving 
blocks bootstrap, where the blocks are row indices. It's easy to build a 
function that allows me to perform a bootstrap on colMeans, but I'm 
flummoxed at the moving blocks bootstrap step.

I could simply hack together a function that, when given block bootstrap 
indices, would do the job but chances are that this would be inefficient 
compared to something that may already be built in. I have examined most 
(probably not all) of the packages that support bootstrap resampling 
statistics, in particular boot and bootstrap, but have come up empty 
handed.

I'm open to suggestions about how best to proceed from others with more 
experience.

Kim Elmore