State-Year Panel for Merging G-W Data (Correlates of War)
cw_gw_panel.RdThis a state-year panel in which the Correlates of War state system is the population of interest. They are matched, as well as one can, with their corollaries in the Gleditsch-Ward system. Its primary use is merging in data demarcated in Gleditsch-Ward state system codes when the primary system in use is the Correlates of War system.
Usage
data(cw_gw_panel)Format
A data frame with the following 6 variables.
stateabbthe state abbreviation, which was the greatest source of agreement between both data sets
yeara numeric vector for the year
gwcodea Gleditsch-Ward state code
ccodea Correlates of War state code
gw_namethe state name as it appears in the Gleditsch-Ward data.
cw_namethe state name as it appears in the Correlates of War data.
Details
The data-raw/ directory on Github contains more information about how
these data were created. The code itself is derived from what peacesciencer
did for its cow_gw_years data. It amounts to the creation of daily data for
both systems before doing a "full join" on where there is the least friction:
state abbreviations. This at least requires the least amount of clean-up.
Use of these data will merge only on the state code and year. The state abbreviations and state names are there for background information, where necessary/appropriate.
peacesciencer's documentation cautions that the differences between the two systems are obvious, if often overstated. Merging one into the other, where possible, will be unproblematic in almost all cases. The biggest headaches concern German unification, Yemeni unification, and the overall history of Serbia/Yugoslavia.
Gleditsch-Ward country names for Württemberg, São Tomé and Príncipe, and Côte d'Ivoire, have manual fixes communicating what the raw data wanted to communicate in ISO-8859-1 (Latin-1) encoding. Mayeul Kauffmann raised this issue on Github, and it's an easy fix, but it's worth reiterating that this fix is more cosmetic or aesthetic than it is practical or functional. You should not ever lean on a country name for serious data management, and the admitted gaudiness of this encoding issue is at most an eyesore in the original data.
Examples
str(cw_gw_panel)
#> tibble [17,720 × 6] (S3: tbl_df/tbl/data.frame)
#> $ stateabb: chr [1:17720] "USA" "USA" "USA" "USA" ...
#> $ year : num [1:17720] 1816 1817 1818 1819 1820 ...
#> $ gwcode : num [1:17720] 2 2 2 2 2 2 2 2 2 2 ...
#> $ ccode : num [1:17720] 2 2 2 2 2 2 2 2 2 2 ...
#> $ gw_name : chr [1:17720] "United States of America" "United States of America" "United States of America" "United States of America" ...
#> $ cw_name : chr [1:17720] "United States of America" "United States of America" "United States of America" "United States of America" ...
head(cw_gw_panel)
#> # A tibble: 6 × 6
#> stateabb year gwcode ccode gw_name cw_name
#> <chr> <dbl> <dbl> <dbl> <chr> <chr>
#> 1 USA 1816 2 2 United States of America United States of America
#> 2 USA 1817 2 2 United States of America United States of America
#> 3 USA 1818 2 2 United States of America United States of America
#> 4 USA 1819 2 2 United States of America United States of America
#> 5 USA 1820 2 2 United States of America United States of America
#> 6 USA 1821 2 2 United States of America United States of America