... Software¹

The current version of this software is available at http://gking.harvard.edu/zelig/, free of charge and open-source (under the terms of the GNU GPL, v. 2).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... Imai²

Assistant Professor, Department of Politics, Princeton University (Corwin Hall, Department of Politics, Princeton University, Princeton NJ 08544; http://Imai.Princeton.Edu, KImai@Princeton.Edu).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... King³

Albert J. Weatherhead III University Professor, Harvard University (Institute for Quantitative Social Sciences, 1737 Cambridge Street, Harvard University, Cambridge MA 02138; http://GKing.Harvard.Edu, King@Harvard.Edu, (617) 495-2027).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... Lau⁴

Ph.D. Candidate, Department of Government, Harvard University (1737 Cambridge Street, Cambridge MA 02138; http://www.people.fas.harvard.edu/~olau, OLau@Fas.Harvard.Edu).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...Zelig^1.1

Zelig is named after a Woody Allen movie about a man who had the strange ability to become the physical reflection of anyone he met -- Scottish, African-American, Indian, Chinese, thin, obese, medical doctor, Hassidic rabbi, anything -- and thus to fit well in any situation.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... charge.^1.2

As specified in the GNU General License v. 2 http://www.gnu.org/copyleft.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... .^2.1

Note that when updating R to the latest release, the installer does not delete previous versions from your C: 1#1 Program Files 1#1 R 1#1 directory. In this example, the subdirectory 1#1 R-2.5.1 1#1 stores R version 2.5.1. Thus, if you have a different version of R installed, you should change the last part of the R program directory file path accordingly.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... ...).^4.1

Multiple imputation is a method of dealing with missing values in your data which is more powerful than the usual list-wise deletion approach. You can create multiply imputed datasets with a program such as Amelia; see King, Honaker, Joseph, Scheve (2000).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... values:^4.2

If you allow more than one variable to vary at a time, you risk confounding the predictive effect of the variables in question.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... details.^4.3

If you choose to create your own bootfn, it must include the the following three arguments: data, the original data frame; one of the sampling methods described in help(boot); and object, the original zelig() output object. The alternative bootstrapping function must sample the data, fit the model, and extract the model-specific parameters.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... output.^4.4

The identical() command checks that numeric values are identical to the maximum number of decimal places (usually 16), and also checks that the the two objects have the same class (numeric, character, integer, logical, or factor). Refer to help(identical) for more information.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... case)^8.1

Why 5? In addition to the intercept term (a variable which is the same in either equation, and so counts only as one variable), the unique variables are 112#112 , 113#113 , 131#131 , and 132#132 .

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... score)^11.1

The linear propensity scores are obtained by transforming back onto a linear scale.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... elements:^11.2

When inapplicable or unnecessary, these elements may equal NULL. For example, when exact matching, match.matrix = NULL.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.