Jed Stephens 31 March 2020
ExpertChoice has two vignettes to help you get started.
The Theoretical Introduction to ExpertChoice focusses on
the theory for designing efficiently for experiments, conjoint and
discrete choice. The Practical Introduction to ExpertChoice
aims to explain how to use this R package. It also gives two worked
examples. The documents reflect each other. For more detail keep
reading.
The need for the ExpertChoice R package emerged from the
methodological desire to implement a discrete choice experiment in my
research. There exists a lack of comprehensive open source software to
assist in the design of discrete choice experiments. Currently there are
three R packages on CRAN that have some overlap with
ExpertChoice: choiceDes, idefix and support.CEs.
Two of these packages are no longer under active development and some of
the functions have not been maintained and consequently no longer work.
Two packages also lack documentation making it difficult for all but
experts in this field to use. ExpertChoice provides a
unified framework suitable for a first time learner to understand how to
design an experiment and convert this experiment into a discrete choice.
Its scope is also wider and more current than the above alternate
packages.
Theoretical introduction to ExpertChoice is the first
vignette: its objective is to explain the theory of experimental design
and discrete choice design. It focusses on explaining how efficiently
measure tests play an important role in the designing process. The
silver object choice experiment, analysed in my dissertation, is one of
the two examples in this vignette. A hypothetical choice experiment on a
restaurant is another.
The second vignette, Practical introduction to
ExpertChoice, provides a worked example of both
experimental designs. The worked examples make it clear how this
procedure could be adapted for the reader’s own experiment. Some of the
more advanced functionality of the package is explored in particular
with the restaurant example.
ExpertChoice now provides a unified open source
alternative to many routines previously only available in SAS and
Ngene.