- DESCRIPTION: Clean and construct measures according to manuscript (pp. XXX)
- CREATES: Cleaned Data for Analysis
- DEPENDS: Raw Manifesto Project Data, Raw ParlGov Cabinet Data, Raw European Representative Democracy Data, Raw Opinion Poll Data
Load the required packages and source the auxiliary functions from src/lib/functions.R:
source("src/lib/functions.R")All raw data files are publicly accessible.
#Load & Tidy Comparative Manifesto Database
cmp <- read_csv("data/raw/MPDataset_MPDS2019b.csv")
cmp <- cmp %>%
filter(country == 42 | country == 21 | country == 41 |
country == 53 | country == 22 | country == 12 |
country == 14 | country == 11 | country == 13 ) %>% #Select Countries under Study
mutate( #Calculate Issue Positions
issue1 = ((per401 + per402 + per407 + per414 + per702) -
(per404 + per405 + per406 + per409 + per412 + per413 + per415 + per701 + per403)),#pos economy
issue2 = (per504 + per506) - (per505 + per507),#pos welfare
issue3 = per108 - per110,#pos welfare
issue4 = per607 - per608,#pos multiculturalism
issue5 = ((per103 + per105 + per106 + per107) -
(per104 + per109)),#pos international affairs
issue6 = per101 - per102,#pos specialrelations
issue7 = per203 - per204,#pos constitution
issue8 = per301 - per302,#pos decentralization
issue9 = (per601 + per603) - (per602 + per604),#pos traditional
issue10 = per201 + per202,#sal democracy
issue11 = per503 + per705 + per706,#sal groups
issue12 = per303 + per304 + per305,#sal governance
issue13 = per408 + per411 + per416,#sal economicgrowth
issue14 = per502,#sal culture
issue15 = per605,#sal law_order
issue16 = per606,#sal harmony
issue17 = per703,#sal farmers
issue18 = per704,#sal middleclass
issue19 = per416 + per501,#sal environment
electionid = interaction(country, edate, drop=TRUE)
) %>%
select(party, country, edate, electionid, rile, issue1, issue2, issue3, issue4, issue5, issue6, issue7, issue8,
issue9, issue10, issue11, issue12, issue13, issue14, issue15,
issue16, issue17, issue18, issue19) #Select all relevant variables
cmp <- calculate_sum_difs(cmp)# Create overlap for each pair of parties
cmp <- cmp %>%
mutate(country = ifelse(substr(electionid_ext,1,2)=="11","Sweden",#Recode country variable
ifelse(substr(electionid_ext,1,2)=="12","Norway",
ifelse(substr(electionid_ext,1,2)=="13","Denmark",
ifelse(substr(electionid_ext,1,2)=="14","Finland",
ifelse(substr(electionid_ext,1,2)=="22","Netherlands",
ifelse(substr(electionid_ext,1,2)=="21","Belgium",
ifelse(substr(electionid_ext,1,2)=="41","Germany",
ifelse(substr(electionid_ext,1,2)=="42","Austria",
ifelse(substr(electionid_ext,1,2)=="53","Ireland",NA))))))))),
electiondate = as.Date(str_sub(electionid_ext,4,13),format="%d/%m/%Y"),
id2 = paste(electionid_ext, party1, party2, sep="."),
party1=recode(party1,#Recode merger parties
`21914` = 21917,
`21221` = 21321,
`23111` = 23113,
`23112` = 23113,
`41112` = 41111,
`41113` = 41111,
`41223` = 41221,
`41222` = 41221,
`21221` = 21321),
party2=recode(party2,#Recode merger parties
`21914` = 21917,
`21221` = 21321,
`23111` = 23113,
`23112` = 23113,
`41112` = 41111,
`41113` = 41111,
`41223` = 41221,
`41222` = 41221,
`21221` = 21321),
id = paste(country,substr(electiondate,1,4),party1, sep="-")) %>%
select(country, electiondate, electionid_ext, id, id2, party1, party2, sum_difs,
rile_party1, rile_party2, rile_difs)#Load & Tidy poll data from Van Der Velden (2015), Jennings and Wlezien (2014) and Askham-Christensen (2012)
load("data/raw/polldata_combined.RData")
polls <- data %>%
pivot_longer(cols = `party 1`:`party 42`,
names_to = "party",
values_to = "polls") %>%
filter(country == "Austria" | country == "Belgium" | country == "Denmark" |
country == "Finland" | country == "Germany" | country == "Ireland" |
country == "Netherlands" | country == "Norway" | country == "Sweden") %>%
mutate(year = as.character(year),
month = as.character(month),
month = ifelse(nchar(month)==1, paste0("0",month), month),
date = paste("01", month, year, sep="/"),
date = as.Date(date, format = "%d/%m/%Y"))
#add cmp code names to party
party <- row.names(cmpcode)
for(i in 1:length(party)){
for(j in 1:length(unique(polls$country))){
select <- which(polls$party==row.names(cmpcode)[i] & polls$country==unique(polls$country)[j])
tmp <- as.character(unique(polls$country)[j])
polls$party[select] <- cmpcode[row.names(cmpcode)[i], tmp]
}
}
polls <- polls %>%
mutate(party = as.numeric(party),
party = recode(party,#Recode merger parties
`21914` = 21917,
`21221` = 21321,
`23111` = 23113,
`23112` = 23113,
`41112` = 41111,
`41113` = 41111,
`41223` = 41221,
`41222` = 41221,
`21221` = 21321),
country_code = substr(party,1,2),
id = paste(country, date, sep=".")) %>%
drop_na(party) %>%
distinct_all() %>%
add_column(event = NA) %>%
arrange(country_code)
#add election dates
id <- read_csv(file = "data/raw/election_id.csv")
for(i in 1:dim(id)[1]){
if(i==1){
start <- i
}
else{
start <- max(which(polls$id==id$id[(i-1)]))+1
}
end <- max(which(polls$id==id$id[i]))
polls$event[start:end] <- str_split(id$id[i], "\\.")[[1]][2]
}
polls <- polls %>%
drop_na(event) %>%
mutate(event = as.Date(event),
month_teller = abs(event-date),#add teller to select 6 months before election
month_teller = round(as.numeric(month_teller)/365 *12,0),
event = paste(country, event, sep = ".")) %>%
add_column(mean_polls = 0,
mean_polls2 = 0) %>%
drop_na(polls)
#Average polls of 6 months prior to election
for(i in 1: length(unique(polls$event))){
tmp <- unique(polls$event)
party <- unique(polls$party[which(polls$event==tmp[i])])
for(j in 1: length(party)){
tmpp <- polls[which(polls$event==tmp[i] & polls$month_teller < 7 & polls$month_teller > 0 & polls$party==party[j]), "polls"]
if(dim(tmpp)[1]>0){
polls$mean_polls[which(polls$event==tmp[i])] <- round(sum(tmpp, na.rm=T)/(6 - length(which(is.na(tmpp)))),2)}
}
}
#Average polls of 12 months prior to election
for(i in 1: length(unique(polls$event))){
tmp <- unique(polls$event)
party <- unique(polls$party[which(polls$event==tmp[i])])
for(j in 1: length(party)){
tmpp <- polls[which(polls$event==tmp[i] & polls$month_teller < 13 & polls$month_teller > 0 & polls$party==party[j]), "polls"]
if(dim(tmpp)[1]>0){
polls$mean_polls2[which(polls$event==tmp[i])] <- round(sum(tmpp, na.rm=T)/(12 - length(which(is.na(tmpp)))),2)}
}
}
polls <- polls %>%
filter(id == event) %>%
mutate(id = paste(country,year,party, sep="-")) %>%
select(id, mean_polls, mean_polls2)Integrate Raw ParlGov Cabinet Data, Raw European Representative Democracy Data, and Termination Cause Data
#Load & tidy Parlgov Data and with Coalition Data Set of Bergmann et al.
parlgov <- read_csv(file = "data/raw/view_cabinet.csv") %>%
filter(country_name == "Austria" | country_name == "Belgium" | country_name == "Denmark" |
country_name == "Finland" | country_name == "Germany" | country_name == "Ireland" |
country_name == "Netherlands" | country_name == "Norway" | country_name == "Sweden") %>%
select(country_name_short, election_date, start_date, cabinet_name, caretaker, cabinet_party, prime_minister,
seats, election_seats_total, party_name_short, party_name_english, party = party_id) %>%
mutate(party = recode(party,#Recode merger parties
`1113` = 1029,
`1487` = 1029,
`255` = 772),
cabinet_name = ifelse(start_date=="1932-05-20", "Dollfuss",
ifelse(start_date=="1920-03-29", "Moeller I",
ifelse(start_date=="1928-06-29", "Moeller II",
ifelse(start_date=="1930-03-30", "Breuning I",
ifelse(start_date=="1930-10-16", "Breuning II",
ifelse(start_date=="1930-12-05", "Breuning III",
ifelse(start_date=="1931-10-10", "Breuning IV",
ifelse(start_date=="1932-12-14", "Kivimaki I",
ifelse(start_date=="1933-07-01", "Kivimaki II", cabinet_name))))))))),
id3 = paste(party_name_short,country_name_short, sep="."))
#add cmp code to parlgov
ids <- read_csv(file = "data/raw/view_party.csv") %>%
mutate(id3 = paste(party_name_short, country_name_short, sep="."),
cmp = recode(cmp,#Recode merger parties
`21914` = 21917,
`21221` = 21321,
`23111` = 23113,
`23112` = 23113,
`41112` = 41111,
`41113` = 41111,
`41223` = 41221,
`41222` = 41221,
`21221` = 21321))%>%
select(id3, cmp_id = cmp) %>%
drop_na(cmp_id) %>%
distinct_all()
parlgov <- full_join(parlgov, ids, by="id3") %>%
drop_na(cmp_id) %>%
distinct_all() %>%
drop_na(country_name_short)
coalitiondata <- read_delim(file = "data/raw/Bergmann_Muller_Strom_Cabinets-Dataset.csv", delim = ";") %>%
filter(v001e == 1 | v001e==2 | v001e==3 | v001e==4| v001e==6 | v001e==9 |
v001e == 12| v001e==13| v001e==16) %>%
select(cabinet_name = v003e, no_cabinetparties = v320e,
govtype = v329e, bargaining_power = v310e ) %>%
mutate(cabinet_name = as.character(cabinet_name),
cabinet_name = recode(cabinet_name,
"Hansson" = "Hansson I",
"F\xe4lldin I" = "Falldin I",
"F\xe4lldin II" = "Falldin II",
"F\xe4lldin III" = "Falldin III",
"Bondevik" = "Bondevik I",
"de Valera VI" = "Valera VI",
"de Valera VII" = "Valera VII",
"de Valera VIII" = "Valera VIII",
"FitzGerald II"="Fitzgerald II",
"Schroeder" = "Schroeder I",
"Kohl V"="Kohl IV",
"Kohl VI"="Kohl V",
"Lipponen" = "Lipponen I",
"von Fieandt" = "Fieandt",
"T\xf6rngren" = "Torngren",
"Karjalainen Iia" = "Karjalainen II",
"Karjalainen Iib" = "Karjalainen III",
"Sukselainen Ib"= "Sukselainen II",
"Sukselainen Ic"= "Sukselainen I",
"Sukselainen IV"= "Miettunen I",
"Holkeri a"="Holkeri I",
"Holkeri b"="Holkeri II",
"Aho a"="Aho I",
"Aho b"="Aho II",
"Paasikivi III"="Paasikivi II",
"Rasmussen IV" = "Rasmussen N IV",
"Rasmussen I" = "Rasmussen N I",
"Rasmussen III" = "Rasmussen N III",
"Rasmussen II" = "Rasmussen N II",
"Schl\xfcter I" = "Schluter I",
"Schl\xfcter II" = "Schluter II",
"Schl\xfcter III" = "Schluter III",
"Schl\xfcter IV" = "Schluter IV",
"Schl\xfcter V" = "Schluter V",
"Spaak" = "Spaak I",
"Dehaenen I"="Dehaene I",
"Van den Boeynants I"="Vanden Boeynants I",
"Van den Boeynants II"="Vanden Boeynants II",
"Tindemans"="Tindemans I",
"Eyskens" = "Eyskens G I",
"Eyskens II" = "Eyskens G II",
"Eyskens III"="Eyskens G III",
"Eyskens V"="Eyskens G V",
"Eyskens IV"="Eyskens G IV",
"M Eyskens" = "Eyskens M",
"Van Houtte" = "Houtte",
"Klima" = "Klima I",
"Vranizky V"="Vranitzky V"))
cabinet_data <- left_join(x = parlgov, y = coalitiondata, by="cabinet_name") %>%
mutate(start_date = as.Date(start_date),
election_date = as.Date(election_date),
end_date = paste(substr(start_date,1,4), substr(start_date,6,7), substr(start_date,9,10), sep="-"),
id3 = paste(id3, cabinet_name, sep="."))
#Add termination causes based on Bergmann, Mueller & Strom book
#Strøm, Kaare; Müller, Wolfgang C. and Bergman, Torbjörn eds. (2008). Cabinets and Coalition Bargaining: The Democratic Life Cycle in Western Europe. Oxford: Oxford University Press. (Paperback 2010).
termination <- read_csv(file = "data/raw/Termination_causes.csv") %>%
select(cabinet_name = v003x, termination_cause = v217y)
cabinet_data <- left_join(x = cabinet_data, y = termination, by = "cabinet_name")
cabinet_data <- slide(data = cabinet_data, Var = "cabinet_name", TimeVar = "start_date", GroupVar = "cmp_id", NewVar = "cabinet_name1")
cabinet_data <- slide(data = cabinet_data, Var = "caretaker", TimeVar = "start_date", GroupVar = "cmp_id", NewVar = "caretaker1")
cabinet_data <- slide(data = cabinet_data, Var = "cabinet_party", TimeVar = "start_date", GroupVar = "cmp_id", NewVar = "cabinet_party1")
cabinet_data <- slide(data = cabinet_data, Var = "prime_minister", TimeVar = "start_date", GroupVar = "cmp_id", NewVar = "prime_minister1")
cabinet_data <- slide(data = cabinet_data, Var = "seats", TimeVar = "start_date", GroupVar = "cmp_id", NewVar = "seats1")
cabinet_data <- slide(data = cabinet_data, Var = "election_seats_total", TimeVar = "start_date", GroupVar = "cmp_id", NewVar = "election_seats_total1")
cabinet_data <- slide(data = cabinet_data, Var = "termination_cause", TimeVar = "start_date", GroupVar = "cmp_id", NewVar = "termination_cause1")
cabinet_data <- slide(data = cabinet_data, Var = "no_cabinetparties", TimeVar = "start_date", GroupVar = "cmp_id", NewVar = "no_cabinetparties1")
cabinet_data <- slide(data = cabinet_data, Var = "govtype", TimeVar = "start_date", GroupVar = "cmp_id", NewVar = "govtype1")
cabinet_data <- slide(data = cabinet_data, Var = "end_date", TimeVar = "start_date", GroupVar = "cmp_id", NewVar = "end_date1", slideBy = 1)
cabinet_data <- cabinet_data %>%
mutate(match_id = paste(election_date, cmp_id, sep="."),
end_date = as.Date(end_date1),
duration_months = round(as.numeric((end_date - start_date))/365*12,0)) %>%
select(match_id, country_name = country_name_short, election_date, start_date, end_date, duration_months, cmp_id, party,
cabinet_party=cabinet_party1, prime_minister=prime_minister1, cabinet_name=cabinet_name1,
no_cabinetparties=no_cabinetparties1, caretaker=caretaker1, govtype=govtype1,
termination_cause=termination_cause1, seats=seats1, tot_seats = election_seats_total1)%>%
add_column(tot_months = 0)
party <- unique(cabinet_data$cmp_id)
for(i in 1:length(party)){
select <- which(cabinet_data$cmp_id==party[i])
if(length(select)==1){
cabinet_data$tot_months[select] = cabinet_data$duration_months[select]
}
if(length(select)>1){
cabinet_data$tot_months[select] <- cumsum(na.omit(cabinet_data$duration_months[select]))
}
}cmp <- cmp %>%
left_join(y = polls, by = "id") %>%
select(country:electionid_ext, id2, party1:rile_difs, polls_party1 = mean_polls, polls2_party1 = mean_polls2) %>%
mutate(id = paste(country,substr(electiondate,1,4),party2, sep="-"))
cmp <- left_join(x = cmp, y = polls, by = "id")
cmp <- cmp %>%
mutate(pcombi1 = paste0(party1, party2),
pcombi2 = paste0(party2, party1),
unique_pcombi = as.numeric(pcombi1==pcombi2)) %>%
filter(unique_pcombi==0) %>%
select(country:party2, pcombi=pcombi1, sum_difs:polls2_party1, polls_party2 = mean_polls, polls2_party2 = mean_polls2) %>%
mutate(match_id = paste(electiondate, party1, sep="."))
df <- left_join(cmp, cabinet_data, by = "match_id") %>%
mutate(match_id = paste(electiondate, party2, sep=".")) %>%
select(match_id, country:electiondate, start_date,
electionid_ext, party = party1, partner = party2, pcombi, sum_difs,
rile_party = rile_party1, rile_partner=rile_party2, rile_difs, polls_party = polls_party1,
polls_partner = polls_party2, polls2_party = polls2_party1, polls2_partner = polls2_party2,
cabinet_party, pm_party = prime_minister, seats_party = seats)
df <- left_join(df, cabinet_data, by = "match_id") %>%
select(country:electiondate, start_date = start_date.x, end_date, electionid_ext,
party = party.x, partner:polls2_partner, cabinet_party = cabinet_party.x,
cabinet_partner = cabinet_party.y, pm_party, pm_partner = prime_minister, seats_party,
seats_partner = seats, tot_seats, cabinet_name:termination_cause, duration_months, tot_months) %>%
mutate(first_election = ifelse(electionid_ext=="11.17/09/1944",1,
ifelse(electionid_ext=="12.08/10/1945",1,
ifelse(electionid_ext=="13.30/10/1945",1,
ifelse(electionid_ext=="14.18/03/1945",1,
ifelse(electionid_ext=="21.17/02/1946",1,
ifelse(electionid_ext=="22.17/05/1946",1,
ifelse(electionid_ext=="41.14/08/1949",1,
ifelse(df$electionid_ext=="42.09/10/1949",1,
ifelse(electionid_ext=="53.04/02/1948",1,0))))))))),
termination_cause = replace_na(termination_cause, 0),
unique_id = paste(electionid_ext, pcombi, sep=".")) %>%
filter(!duplicated(unique_id)) %>%
select(-unique_id) %>%
filter(first_election==0) %>%
select(-first_election)# Mutate data
df <- slide(data = df, Var = "sum_difs", TimeVar = "electiondate", GroupVar = "pcombi", NewVar = "l_sum_difs")
df <- df %>%
mutate(pcombi = as.numeric(pcombi),
year = substr(electiondate, 1,4),
year = as.numeric(year),
d_issue_distance = sum_difs-l_sum_difs,
cabinet_pair = cabinet_party + cabinet_partner,
conflict = recode(termination_cause, `2`= 0),
sum_polls = polls_party + polls_partner,#and voluntary early & regular end of term are 1 category
popularity = ifelse(sum_polls==0, NA, log((polls_party + polls_partner)/(seats_party + seats_partner))),
popularity2 = ifelse(sum_polls==0, NA, log((polls2_party + polls2_partner)/(seats_party + seats_partner))),
rile = ifelse(rile_party<0 & rile_partner<0,0,
ifelse(rile_party>0 & rile_partner>0,0,1))) %>%#rile: being on other side = 1, same side = 0
add_column(rile_mean = 0,
rile_median = 0)
#calculate mean and median per election cycle
cutoffs_rile <- df %>%
group_by(electionid_ext) %>%
summarise(rile_mean = mean(rile_party + rile_partner, na.rm = T)/2,
rile_median = median(rile_party + rile_partner, na.rm = T)/2)
for(i in 1:dim(cutoffs_rile)[1]){
df$rile_mean[df$electionid_ext==cutoffs_rile$electionid_ext[i]] <- cutoffs_rile$rile_mean[i]
df$rile_median[df$electionid_ext==cutoffs_rile$electionid_ext[i]] <- cutoffs_rile$rile_median[i]
}
df <- df %>%
mutate(rile_mean = ifelse(rile_party<rile_mean & rile_partner<rile_mean,0,
ifelse(rile_party>rile_mean & rile_partner>rile_mean,0,1)),
rile_median = ifelse(rile_party<rile_median & rile_partner<rile_median,0,
ifelse(rile_party>rile_median & rile_partner>rile_median,0,1)))
#calculate experience
tmp <- tibble(pairs=unique(df$pcombi),
select = 0)
for(i in 1:length(pairs)){
tmp$select[i] <- ifelse(length(which(df$pcombi==tmp$pairs[i] & df$cabinet_pair==1))>0,1,0)
}
select <- tmp$pairs[which(tmp$select==1)]
tmp <- df[,c("duration_months", "pcombi", "cabinet_name", "start_date")]
tmp <- slide(tmp, Var = "duration_months", TimeVar = "start_date", GroupVar = "pcombi", NewVar = "l_duration")
tmp <- tmp %>%
mutate(l_duration = replace_na(l_duration,0),
experience = duration_months + l_duration,
match_id = paste(cabinet_name, pcombi, sep=".")) %>%
select(match_id, experience)
df <- df%>%
mutate(match_id = paste(cabinet_name, pcombi, sep=".")) %>%
left_join(y=tmp, by="match_id") %>%
drop_na(start_date) %>%
select(match_id, country:rile_median,
experience) %>%
mutate(experience = replace_na(experience, 0),
experience = round(experience/tot_months,2))
#add controls
controls <- read_csv(file = "data/raw/controls.csv") %>%
mutate(id = paste(counrty, year, sep = "."),
type_conflict = replace_na(type_conflict,0)) %>%
select(enps:id)
df <- df %>%
mutate(id = paste(country, year, sep=".")) %>%
left_join(y = controls, by = "id") %>%
mutate(year = ifelse(electionid_ext=="13.22/09/1953", 1954,# For tscs reasons, double election years can't be dealed with
ifelse(electionid_ext=="53.24/11/1982", 1983,# (Denemarken 1953, iceland 1969 en Ierland 1982)
ifelse(electionid_ext=="15.25/10/1959", 1960, year))),# therefore 2nd election is coded as year later
pec_bergman = ifelse(country=="Norway" & year == 1965, 1,
ifelse(country=="Norway" & year == 1981, 1,
ifelse(country=="Norway" & year == 1989, 1,
ifelse(country=="Norway" & year == 2001, 1,
ifelse(country=="Sweden" & year == 2006, 1,
ifelse(country=="Ireland" & year == 2007, 1, 0))))))) #Control for pre-coalition agreements Bergman
#Control for pre-coalition agreements Golder's BJPS 2006
pec <- read_csv(file = "data/raw/PEC_Golder2006.csv") %>%
mutate(id = paste(country, year, sep=".")) %>%
select(pec_golder = pec, id)
df <- df %>%
left_join(y = pec, by = "id") %>%
filter(cabinet_pair>0) %>%
mutate(unique_id = paste(electionid_ext, pcombi, sep="."),
cabinet_pair = recode(cabinet_pair,
`1`=0,
`2`=1)) %>%
filter(!duplicated(unique_id)) %>%
select(country:electionid_ext, year, party:pcombi, sum_difs, l_sum_difs,
d_issue_distance, seats_party, seats_partner,tot_seats, polls_party:polls2_partner,
popularity,popularity2, cabinet_name:termination_cause, conflict, cabinet_pair, duration_months,
tot_months, experience, rile_party:rile_difs, rile:rile_median, enps:pec_golder) %>%
drop_na(popularity)
#Save Dataset
save(df, file="data/intermediate/cleaned_dyadic_data.RData")
#Check Correlations
# as a default this function outputs a correlation matrix plot
df%>%
select(d_issue_distance, experience, popularity,
conflict, no_cabinetparties,
rile, enps, gdp) %>%
ggstatsplot::ggcorrmat(
type = "robust", # correlation method
sig.level = 0.05, # threshold of significance
p.adjust.method = "holm", # p-value adjustment method for multiple comparisons
cor.vars = c(d_issue_distance:gdp), # a range of variables can be selected
cor.vars.names = c(
"Δ Issue Distance", # variable names
"Experience", # variable names
"Popularity",
"Conflict",
"# Coalition Parties",
"Ideological Position",
"ENPS",
"GDP"),
matrix.type = "upper", # type of visualization matrix
colors = c("#B2182B", "white", "darkgreen"),
title = "Correlalogram for Variables under Study",
lab_size = 2
)
## Dependent Variable
df %>%
ggplot(aes(x = d_issue_distance)) +
geom_histogram(fill = "gray85", colour = "black", binwidth = 10) +
theme_classic() +
theme(plot.title = element_text(hjust = 0.5)) +
labs(x = "Range: -73.1 - 71.4", y="", title = "Δ Issue Distance \n Mean: -1.44, Standard Deviation: 20.02")
- Popularity (H1a)
- Experience (H1b)
- Conflict(H1c)
- Being a Government Proto-Coalition (H1)
- Number of Coalition Parties (control variable)
- Effective Number of Parties (control variable)
- % Change GDP Growth (control variable)
- Being on the Same Ideological Side (control variable)
## Independent Variables
rbind(tibble(
freq = c(round(df$popularity,2), round(df$experience, 2), df$no_cabinetparties,
round(df$enps,2), round(df$gdp,2)),
id = c(rep("Popularity \n Mean: -2.15, Standard Deviation: 1.48 \n Range: -8.13 - 1.72", length(round(df$popularity,2))),
rep("Experience \n Mean: 0.16, Standard Deviation: 0.19 \n Range: 0 - 1", length(round(df$experience, 2))),
rep("Number of Coalition Parties \n Mean: 2.53, Standard Deviation: 1.08 \n Range: 1 - 5", length(round(df$no_cabinetparties, 2))),
rep("Effective Number of Parties (ENPS) \n Mean: 74.61, Standard Deviation: 7.04 \n Range: 57.75 - 88.14", length(round(df$enps, 2))),
rep("% Change GDP Growth \n Mean: 1.91, Standard Deviation: 2.65 \n Range: -4.98 - 6.42", length(round(df$gdp, 2)))))) %>%
ggplot(aes(x = freq)) +
geom_histogram(fill = "gray85", colour = "black") +
theme_classic() +
theme(plot.title = element_text(hjust = 0.5)) +
facet_wrap(~ id, ncol = 3,scales = "free") +
labs(x="",y="", title = "Continues Independent Variables")
rbind(tibble(
value = c("No Conflict","Conflict","Not a Government Proto-Coalition",
"Government Proto-Coalition","Same Ideological Side",
"Opposite Side"),
freq = c(table(df$conflict2),table(df$cabinet_pair2),table(df$rile2)),
id = c(rep("Conflict \n Median: No Conflict", 2),
rep("Being a Government Proto-Coalition \n Median: Not a Government Proto-Coalition", 2),
rep("Being on the Same Ideological Side \n Median: Same Ideological Side", 2)))) %>%
ggplot(aes(x = value, y = freq)) +
geom_col(fill = "gray85", colour = "black" ) +
theme_classic() +
theme(plot.title = element_text(hjust = 0.5)) +
facet_wrap(~ id, ncol = 3,scales = "free") +
labs(x="",y="", title = "Dichotomous Independent Variables")
