Extract the information in a roll call matrix as a series of vectors with voting decision, a unique identifier for the legislator and a unique identifier for the roll call.
vectorRepresentation(object, dropList = list(codes = c("missing", "notInLegis")))A matrix with z rows, where z is the
number of non-missing entries in object$votes, with
‘missingness’ defined by the codes component of the
dropList. The matrix has 3 columns:
the voting decision, either a 1 if the
corresponding element of the roll call matrix object$votes
is in the yea component of object$codes, or a 0
if the corresponding element of the roll call matrix is in the
nay component of object$codes. Non-missing entries of
the roll call matrix are not stored.
the row of the roll call matrix object$votes that
supplied the voting decision; i.e., a unique identifier for the
legislator generating this vote
the column of the roll call matrix object$votes that
supplied the vote; i.e., a unique identifier for the vote.
an object of class rollcall
a dropList; see dropRollCall
Simon Jackman simon.jackman@sydney.edu.au
It is often the case that roll call matrices are sparse, say, when the roll call matrix has an “overlapping generations” structure; e.g., consider forming data by pooling across a long temporal sequence of legislatures such that relatively few of the legislators in the data set actually vote on any given roll call. In such a case, representing the data as a roll call matrix is not particularly helpful nor efficient, either for data summaries or modeling.
rollcall
data(s109)
y <- vectorRepresentation(s109)
apply(y,2,table,exclude=NULL)
Run the code above in your browser using DataLab