Title: SimDesign: An R Package for Executing Monte Carlo Simulation
Experiments

Author: Phil Chalmers

Abstract:
Monte Carlo simulation experiments are useful for studying the behaviour of
statistical estimates and estimators, particularly in situations where formal
statistical theory is intractable or unavailable. At their core, this class of
simulation experiments consist of generating random data given a set of
simulation design and data-generating characteristics, analyzing said data
using one or more inferential methods or competing models, and repeating this
generate-analyse sequence numerous times until suitable information can be
drawn regarding the behaviour of the population sampling distribution. Results
from these analyses are then summarized using suitable meta-statistical
methods and subsequent graphical presentations, often for the purpose of
dissemination. This talk will provide an overview of how the R package
"SimDesign" can be used for designing and streamlining such simulation
experiments. The general structural philosophy of the package will be
discussed, along with the associated implementation, organization, debugging,
and safety features that are desirable when constructing (and reconstructing)
the associated computer code for intensive simulation experiments.