read roll call data in Poole-Rosenthal KH format
Creates a rollcall object from the flat file format for roll call data used by Keith Poole and Howard Rosenthal.
readKH(file, dtl=NULL, yea=c(1,2,3), nay=c(4,5,6), missing=c(7,8,9), notInLegis=0, desc=NULL, debug=FALSE)
file: string, name of a file or URL holding KH datadtl: string, name of a file or URL holding KH dtl file (information about votes); default is NULL, indicating no dtl fileyea: numeric, possibly a vector, code(s) for a Yea vote in the rollcall context (or a correct answer in the educational testing context). Default is c(1,2,3), which corresponds to Yea, Paired Yea, and Announced Yea in Poole/Rosenthal data files.nay: numeric, possibly a vector, code(s) for a Nay vote in the rollcall context (or an incorrect answer in the educational testing context). Default is c(4,5,6), which corresponds to Announced Nay, Paired Nay, and Nay in Poole/Rosenthal data files.missing: numeric and/or NA, possible a vector, code(s) for missing data. Default is c(0,7,8,9,NA); the first four codes correspond to Not Yet a Member, Present (some Congresses), Present (some Congresses), and Not Voting.notInLegis: numeric or NA, possibly a vector, code(s) for the legislator not being in the legislature when a particular roll call was recorded (e.g., deceased, retired, yet to be elected). Default is 0 for Poole/Rosenthal data files.desc: string, describing the data, e.g., 82nd U.S. House of Representatives; default is NULLdebug: logical, print debugging information for net connectionan object of class rollcall, with components created using the identifying information in the Poole/Rosenthal files. If the function can not read the file (e.g., the user specified a URL and the machine is not connected to the Internet), the function fails with an error message (set debug=TRUE to help resolve these issues).
Keith Poole and Howard Rosenthal have gathered an impressive collection of roll call data, spanning every roll call cast in the United States Congress. This effort continues now as a real-time exercise, via a collaboration with Jeff Lewis (109th Congress onwards). Nolan McCarty collaborated on the compilation of roll call data for the 102nd through 108th Congress.
This function relies on some hard-coded features of Poole-Rosenthal flat files, and assumes that the file being supplied has the following structure (variable, start-end columns):
This function reads data files in that format, and creates a rollcall, for which there are useful methods such as summary.rollcall. The legis.data component of the rollcall object is a data.frame which contains:
state: a 2-character string abbreviation of each legislator' stateicpsrState: a 2-digit numeric code for each legislator's state, as used by the Inter-university Consortium for Political and Social Research (ICPSR)cd: numeric, the number of each legislator's congressional district within each state; this is always 0 for members of the SenateicpsrLegis: a unique numeric identifier for each legislator assigned by the ICPSR, as corrected by Poole and Rosenthal.partyName: character string, the name of each legislator's political partyparty: numeric, code for each legislator's political party; see https://legacy.voteview.com/PARTY3.HTMThe rownames attribute of this data frame is a
concatenation of the legislators' names, party abbreviations (for
Democrats and Republicans) and state, and (where appropriate), a
district number; e.g., Bonner (R AL-1). This tag is also
provided in the legis.name component of the returned rollcall
object.
Poole and Rosenthal also make dtl files available for Congresses 1 through 106. These files contain information about the votes themselves, in a multiple-line per vote ascii format, and reside in the dtl director of Poole's web site, e.g., https://legacy.voteview.com/k7ftp/dtl/102s.dtl is the dtl
file for the 102nd Senate. The default is to presume that no such file exists. When a dtl file is available, and is read, the votes.data attribute of the resulting rollcall
object is a data.frame with one record per vote, with the following variables:
date: vector of class Date, date of the rollcall, if available; otherwise NULLdescription: vector of mode character, descriptive textThe dtl files are presumed to have the date of the rollcall in
the first line of text for each roll call, and lines 3 onwards contain
descriptive text.
Finally, note also that the Poole/Rosenthal data sets often include the U.S. President as a pseudo-legislator, adding the announced positions of a president or the administration to the roll call matrix. This adds an extra legislator to the data set and can sometimes produce surprising results (e.g., a U.S. Senate of 101 senators), and a legislator with a surprisingly low party loyalty score (since the President/administration only announces positions on a relatively small fraction of all Congressional roll calls).
Poole, Keith and Howard Rosenthal. 1997. Congress: A Political-Economic History of Roll Call Voting. New York: Oxford University Press.
Poole, Keith. https://legacy.voteview.com
Rosenthal, Howard L. and Keith T. Poole. United States Congressional Roll Call Voting Records, 1789-1990: Reformatted Data [computer file]. 2nd ICPSR release. Pittsburgh, PA: Howard L. Rosenthal and Keith T. Poole, Carnegie Mellon University, Graduate School of Industrial Administration [producers], 1991. Ann Arbor, MI: Inter-university Consortium for Political and Social Research [distributor], 2000. https://www.icpsr.umich.edu/icpsrweb/ICPSR/studies/09822
rollcall
## Not run: h107 <- readKH("https://voteview.com/static/data/out/votes/H107_votes.ord", desc="107th U.S. House of Representatives") s107 <- readKH("https://voteview.com/static/data/out/votes/S107_votes.ord", desc="107th U.S. Senate") ## End(Not run)
Simon Jackman simon.jackman@sydney.edu.au