Version 1.03 of the R package boolean has been uploaded to CRAN and is
now available.
boolean implements partial-observability logit and probit models for
testing Boolean hypotheses. It permits researchers to model the
probability of the occurrence of a given outcome as a complex function
of the probabilities that other outcomes will occur (or other
conditions will be fulfilled). For example, if p(Y) = p(A) * p(B)
(that is, A and B jointly produce Y, and the absence of either
precludes it), the boolean routine models p(A) and p(B) as logit or
probit curves and p(Y) as their product. Similarly, if A or B produces
Y, p(Y) = 1 - [(1-p(A)) * (1-p(B))] becomes the functional form, again
with p(A) and p(B) estimated as logit/probit curves. Arbitrarily
convoluted combinations -- e.g., (A and B) or (C and D) produces Y; or,
p(Y) = 1 - [1-p(A)*p(B)] * [1-p(C)*p(D)] -- can be estimated, and
dependence of probabilities can be modeled by including some of the
same independent variables in the separate logit or probit equations.
The derivation is in Braumoeller, "Causal Complexity and the Study of
Politics," Political Analysis 11(3), 209-233.
Version 1.03 is a service update which ensures that all documentation,
in particular the documentation of S4 methods objects, is compatible
with R 1.8. Previous versions will produce warnings upon compilation.
Bear F. Braumoeller
Assistant Professor
Department of Government
Harvard University
http://www.people.fas.harvard.edu/~bfbraum
_______________________________________________
R-packages mailing list
R-packages at stat.math.ethz.ch
https://www.stat.math.ethz.ch/mailman/listinfo/r-packages
