Thanks to everyone who responded to my query! I got 6 responses, and 5
confirmed that data should be simulated by row, and the 6th suggested an
alternative strategy that would take correlation among variables into
account (but is a little more complex than I need for this project). My
original question was:
>I am trying to program a bootstrap simulation of a data set. When I simulate data, should I: 1) sample each x-variable independently, or 2) select a set of x-variables (i.e. pick a data row and use the set of measurements)? I am guessing that I should go by row, so that I don't disrupt any correlations among the variables. Any thoughts?
>
Brian
--
----------------------------------------------------------
Brian R. Mitchell
Post-Doctoral Associate
University of Vermont
The Rubenstein School of Environment and Natural Resources
81 Carrigan Drive
Burlington, VT 05405-0088
(802) 656-2496
[log in to unmask]
----------------------------------------------------------
|