I am writing to announce the availability of a new package on CRAN:
"boolean," which permits users to run Boolean logit and probit
analyses.
BACKGROUND
The impact of independent probabilistic causal processes is often
thought to cumulate in a manner consistent with Boolean logic --
specifically, a logic in which the atomic elements are connected by
"and" or "or." For example, diet or heredity can lead to
heart
failure, apathy or indifference can lead to nonvoting, plant survival
depends on light and water and proper soil, and so on. This form of
cumulation has been referred to as "causal complexity" or
"multiple
causal paths" to a given (non-)outcome (see e.g. Ragin 1987).
The boolean() command is designed for use in such situations. It
models the probabilities of the antecedent processes ("paths") as
logit
and probit curves, constructs a likelihood function from the logic of
their interaction, and maximizes to obtain parameter estimates. Any
number of causal paths can be modeled, in any combination -- A and B
and C cause Y, A and (B or (C and D)) cause Y, etc. For more details
and the derivation of the procedure see <a
href="http://www.people.fas.harvard.edu/~bfbraum/">Braumoeller
(2003)</a>.
PACKAGE
In addition to performing Boolean logit and probit analyses, the
boolean package contains functions to assist the user in a variety of
ways:
* boolprep() simplifies input for the boolean command
* boolfirst() graphs first differences after estimation
* boolprof() produces profile likelihoods for diagnostics
In addition, the plot() command produces an array of graphs of first
differences or profile likelihoods for all of the relevant coefficients
in the model.
For more information please visit <a
href="http://www.people.fas.harvard.edu/~bfbraum/">http://
www.people.fas.harvard.edu/~bfbraum/</a>
REFERENCES
Braumoeller, Bear F. (2003) "Causal Complexity and the Study of
Politics." Political Analysis 11(3): 209-233.
Ragin, Charles C. (1987) The Comparative Method: Moving Beyond
Qualitative and Quantitative Strategies. Berkeley: University of
California Press.
<A HREF="mailto:bbraumoeller at wcfia.harvard.edu">Bear F.
Braumoeller</A>
Assistant Professor
Department of Government
Harvard University