Zelig: Everyone's Statistical Software
Kosuke Imai, Gary King and Olivia Lau
Version 1.0
(Available at http://gking.harvard.edu/zelig)
A growing proportion of statisticians and methodologists from many
disciplines are converging on R, a powerful statistics package and
programming language. As an open source project, R is freely
accessible. With thousands of contributors who have written hundreds
of packaged routines, R can deal with nearly any statistical problem.
Although this high level of participation may be its greatest
strength, the enormous diversity in approaches to statistical
inference covered by R often results in a virtual babel of competing
functions and inconsistent syntax.
To address these problems from a common perspective, we have created
Zelig, a single, easy-to-use program, with a unified framework and
syntax, that can estimate, help interpret, and present the results of
a large range of statistical methods. It literally _is_ "everyone's
statistical software" because Zelig uses R code from many researchers.
We also hope it will _become_ "everyone's statistical software"
for
applications, and we have designed it so that anyone can use it or add
their methods to it. Zelig comes with detailed, self-contained
documentation that minimizes startup costs for Zelig and R, automates
graphics and summaries for all models, and, with only three simple
commands required, generally makes the power of R accessible for all
users. Zelig also works well for teaching, and is designed so that
scholars can use the same program they use for their research.
Zelig adds considerable infrastructure to improve the use of existing
methods. Zelig
- implements and generalizes the program Clarify (for Stata),
which translates hard-to-interpret coefficients into quantities of
interest;
- combines multiply imputed data sets (such as output from Amelia) to
deal with missing data;
- automates bootstrap simulation for all models;
- uses sophisticated nonparametric matching commands which improve
parametric procedures (via MatchIt);
- allows one-line commands to run analyses in all designated strata;
- automates the creation of replication data files so that you (or,
if you wish, anyone else) can replicate the results of your
analyses (hence satisfying the replication standard); and
- makes conditional population and superpopulation inferences.
If you wish to add your models to Zelig, the documentation also
includes instructions for how to do so. Zelig is easily expandable,
and various contributors are currently working to include their
methods in Zelig. As Zelig grows, you will have access to an
increasing range of methods and models.
