Terra

NEWS.md

@@ -2,6 +2,7 @@
 
 Additions include:
 
+- `african_coups`: a data set on the correlates of African coups from 1960 to 1975 (1982). Useful for replication of an old debate, and for pedagogical instruction about linear models (especially about interactions).
 - `DCE12`: a data set on domestic conflict events in 2012, useful for teaching about regression of count data.
 - `DAPO`: a data set on the determinants of public opinion in seven Arab countries.
 - `EBJ`: a data set on the economic benefits of post-conflict justice institutions.

R/rd-DAPO.R

@@ -14,7 +14,6 @@ NULL
 #' \item{\code{capsub}}{the composite index of national capabilities (capability ratio) of the subject country}
 #' \item{\code{capobj}}{the composite index of national capabilities (capability ratio) of the object country}
 #' \item{\code{securtie}}{a dummy variable indicating at least an informal security tie between the subject and object}
-#' \item{\code{dyadtrde}}{the volume of dyadic trade between subject and object}
 #' \item{\code{export}}{the volume of exports from the subject to the object}
 #' \item{\code{import}}{the volume of imports to the subject from the object}
 #' \item{\code{subgdp}}{the gross domestic product (GDP) of the subject}
@@ -26,8 +25,11 @@ NULL
 #'
 #' Exact coding issues/peculiarities are best addressed by reading the reference
 #' article. To maximally reproduce the article's analyses, the user will need
-#' to create some variables. However, I think this is a learning experience for
-#' students.
+#' to create some variables. The information is here, but you'll need to create
+#' a variable for dyadic trade (and as a percentage of the subject's GDP),
+#' GDP-adjusted imports, a means to filter out Israel from the analysis, and
+#' some of the information reported in Table 1. However, I think this is a
+#' learning experience for students.
 #'
 #' @references
 #'

R/rd-ESS10NO.R

@@ -4,7 +4,7 @@ NULL
 #' Norwegian Attitudes toward European Integration (2021-2022)
 #'
 #' This is a simple data set to illustrate the use of sampling weights from
-#' the European Social Survey.
+#'  the European Social Survey.
 #'
 #' @format A data frame with 1,411 observations on the following 24 variables.
 #' \describe{

R/rd-african_coups.R

@@ -0,0 +1,81 @@
+#' @importFrom tibble tibble
+NULL
+
+#' Modeling Coups in Africa, 1960 to 1975 (1982)
+#'
+#' A data set on modeling coups in Africa using data from the period between
+#'  1960 to 1975 (1982). These data offer a partial replication of Jackman
+#'  (1978).
+#'
+#' @format A data frame with the following 10 variables.
+#' \describe{
+#' \item{\code{iso3c}}{a three-character ISO code for state identification}
+#' \item{\code{country}}{an English country name}
+#' \item{\code{jci}}{Jackman's (1978) coup index from 1960 to 1975}
+#' \item{\code{tmis}}{Johnson et al.'s (1986) total military involvement score}
+#' \item{\code{agperc}}{an estimate of the percentage of the country's labor force in agriculture}
+#' \item{\code{literacy_cnts}}{an estimate of countrywide literacy from around 1965}
+#' \item{\code{literacy_ba}}{another estimate of countrywide literacy from around 1965}
+#' \item{\code{leperc}}{an estimate of the size of the largest ethnic group, as a percentage}
+#' \item{\code{partydom}}{the percentage of the vote received by the largest party in the country prior to independence}
+#' \item{\code{turnout}}{the turnout (as a percentage) at the independence referendum}
+#' }
+#'
+#' @details
+#'
+#' Data exist for instructional purposes, especially about modeling interactions.
+#' Reading Jackman (1978) and Johnson et al. (1986) will provide more information
+#' about the stake of this debate.
+#'
+#' English country names are country names from around the time of publication.
+#' Take note of older names of "Dahomey", "Swaziland", "Upper Volta", and "Zaire."
+#' The three-character ISO codes are current, though it comes with the
+#' acknowledgment that Dahomey used to have a different ISO code.
+#'
+#' Jackman (1978) is deceptively opaque on what he's doing for the ethnic group
+#' variable and arguably misleads on what his turnout variable is actually from.
+#' In the case of the ethnic group variable, it's the difference between saying
+#' the largest ethnic group in Rwanda is 98% of the population versus 80% of
+#' the population. I'm uncertain with what he's doing with what Morrison et al.
+#' (1989) define as "ethnic clusters".
+#'
+#' Mercifully, the coup variables come from a replication by Johnson et al. (1986).
+#' The bulk of this is arguably lost to history.
+#'
+#' Ideally, I'd have Morrison's (1972) \emph{Black Africa}, but I do not. I have
+#' a copy of a 1989 update, though. That's what I consulted in constructing
+#' this data set.
+#'
+#' Related: the agricultural variable is a midway point between columns B and
+#' columns C in Table 3.11 of Morrison et al. (1989). I do not think this is too
+#' far removed from what Jackman was looking at in an older version of the same
+#' data, but there will be slight differences. It's the difference of "these
+#' variables came from 1966" versus "this is an imputation of 1960 to 1970". The
+#' latter is what I offer here.
+#'
+#' The literacy variables have suffices communicating where I obtained them. The
+#' Cross-National Time Series Database has a variable effectively communicating
+#' this information that I was using first. These data come 1965 in that data set.
+#' Jackman and Johnson et al. are assuredly using Morrison's almanac. That
+#' information is in Table 4.11 or Morrison et al. (1989).
+#'
+#' Ethiopia is conspicuously missing in the party dominance variable.
+#'
+#' I am 99.9% sure the turnout variable is Table 5.9 in Morrison et al. (1989).
+#' Jackman (1978) says this is from *before* independence but I'm confident he
+#' meant it was the turnout at the independence referendum.
+#'
+#' @references
+#'
+#' Jackman, Robert W. 1978. "The Predictability of Coups d'etat: A Model with
+#' African Data." \emph{American Political Science Review} 72(4): 1262-75.
+#'
+#' Johnson, Thomas H., Robert O. Slater, and Pat McGowan. 1986. "Explaining
+#' African Military Coups d'Etat, 1960-1982." \emph{American Political Science
+#' Review} 80(4): 225-49.
+#'
+#' Morrison, Donald George, Robert Cameron Mitchell, and John Naber Paden. 1989.
+#' \emph{Black Africa: A Comparative Handbook} (2nd ed.). New York, NY: The Free
+#' Press.
+
+"african_coups"

data-raw/african_coups/african_coups.R

@@ -0,0 +1,48 @@
+CNTS <- readxl::read_excel("/home/steve/Dropbox/data/cnts/CNTSDATA.xls")
+library(tidyverse)
+
+# CNTS %>%
+#   mutate(iso3c = countrycode::countrycode(country, "country.name", "iso3c")) %>%
+#   #select(country, iso3c, year, polit01:polit15) %>%
+#   mutate(iso3c = case_when(
+#     country == "CEN AFR REP" ~ "CAF",
+#     country == "CEN AFR EMP" ~ "CAF",
+#     country == "ETH'PIA FDR" ~ "ETH",
+#     country == "ETH'PIA PDR" ~ "ETH",
+#     TRUE ~ iso3c
+#   )) %>%
+#   mutate(iso2c = countrycode::countrycode(iso3c, "iso3c", "iso2c")) %>%
+#   filter(year %in% c(1964:1966)) %>%
+#   select(country, year, iso2c, iso3c, industry3, school12) %>%
+#   # Take ag from 1966 if you can. Take school from 1965 if you can.
+#   filter(iso3c %in% c("MDG", "GNQ", "GNB",
+#                       "COM", "SYC", "ZWE",
+#                       "ANG", "MOZ", "STP", "BWA",
+#                       "CPV", "DJI", "LSO", "MUS",
+#                       "SWZ"))
+
+
+
+readxl::read_excel("/home/steve/Dropbox/projects/stevedata/data-raw/african_coups/african_coups.xlsx") %>%
+  arrange(iso3c) %>%
+  select(iso3c, everything(), -agperc_cnts) %>%
+  rename(agperc = agperc_ba) -> african_coups
+
+
+save(african_coups, file="data/african_coups.rda")
+
+#
+#   mutate(ddd = 100 - agperc_ba) %>%
+#   mutate(m = ddd+literacy_ba,
+#          p = ifelse(turnout >= 20, 1, 0),
+#          c = ifelse(leperc >= 44, 1, 0)) -> A
+#
+#
+# broom::tidy(M1 <- lm(jci ~ m*partydom*leperc, A))
+# broom::tidy(M2 <- lm(jci ~ m + c + partydom + p + partydom*p + c*partydom + c*p, A))
+# broom::tidy(M3 <- lm(jci ~ m + c + partydom + p, A))
+
+
+
+sbroom::augment(M1) %>%
+  ggplot(.,aes(.fitted, jci)) + geom_point()