Macro regions
## [1] 71
## # A tibble: 10 × 2
## V2 macroregion
## <dbl> <chr>
## 1 356 India
## 2 156 East Southeast Asia
## 3 344 East Southeast Asia
## 4 392 East Southeast Asia
## 5 410 East Southeast Asia
## 6 158 East Southeast Asia
## 7 458 East Southeast Asia
## 8 608 East Southeast Asia
## 9 764 East Southeast Asia
## 10 360 East Southeast Asia
## # A tibble: 7 × 2
## macroregion n
## <chr> <int>
## 1 Western Europe Offshoots 20499
## 2 East Southeast Asia 13011
## 3 Latin America Caribbean 12589
## 4 Middle East North Africa MENA 12165
## 5 Sub-Saharan Africa 12097
## 6 Eastern Europe Post-Soviet 11613
## 7 India 2001
## macroregion n
## 1 East Southeast Asia 23218
## 2 Western Europe Offshoots 19417
## 3 Latin America Caribbean 17439
## 4 Eastern Europe Post-Soviet 13951
## 5 Middle East North Africa MENA 12321
## 6 Sub-Saharan Africa 4948
## 7 South Asia 4234
## 8 India 1692
## region n
## 1 Eastern Europe Post-Soviet 18410
## 2 Middle East North Africa MENA 16827
## 3 East Southeast Asia 13853
## 4 Latin America Caribbean 13435
## 5 Western Europe Offshoots 11893
## 6 Sub-Saharan Africa 9869
## 7 India 4078
## 8 Pakistan 1200
## macroregion_h n
## 1 Eastern Europe Post-Soviet 18410
## 2 Middle East North Africa MENA 16827
## 3 East Southeast Asia 13853
## 4 Latin America Caribbean 13435
## 5 Western Europe Offshoots 11893
## 6 Sub-Saharan Africa 9869
## 7 India 4078
## 8 South Asia 1200
# W5 (2006): V2 -> macroregion
w5_unmapped <- wvs5 %>%
distinct(V2) %>%
filter(!is.na(V2)) %>%
left_join(macromapw5 %>% distinct(V2), by = "V2") %>%
mutate(mapped = V2 %in% macromapw5$V2) %>%
filter(!mapped)
nrow(w5_unmapped)
## [1] 0
w5_unmapped %>% arrange(V2)
## # A tibble: 0 × 2
## # ℹ 2 variables: V2 <dbl>, mapped <lgl>
# Also check for any respondents with NA macroregion (should be 0 or very small if mapping complete)
wvs5 %>% summarise(n_na_macroregion = sum(is.na(macroregion)))
## # A tibble: 1 × 1
## n_na_macroregion
## <int>
## 1 0
# W6 (2012): V2 -> region
w6_unmapped <- wvs6 %>%
distinct(V2) %>%
filter(!is.na(V2)) %>%
left_join(countrymap_w6 %>% distinct(V2), by = "V2") %>%
mutate(mapped = V2 %in% countrymap_w6$V2) %>%
filter(!mapped)
nrow(w6_unmapped)
## [1] 0
w6_unmapped %>% arrange(V2)
## [1] V2 mapped
## <0 rows> (or 0-length row.names)
wvs6 %>% summarise(n_na_region = sum(is.na(region)))
## n_na_region
## 1 0
# W7 (2022): B_COUNTRY -> macroregion
w7_unmapped <- wvs7 %>%
distinct(B_COUNTRY) %>%
filter(!is.na(B_COUNTRY)) %>%
left_join(macromapw7 %>% distinct(B_COUNTRY), by = "B_COUNTRY") %>%
mutate(mapped = B_COUNTRY %in% macromapw7$B_COUNTRY) %>%
filter(!mapped)
nrow(w7_unmapped)
## [1] 0
w7_unmapped %>% arrange(B_COUNTRY)
## [1] B_COUNTRY mapped
## <0 rows> (or 0-length row.names)
wvs7 %>% summarise(n_na_macroregion = sum(is.na(macroregion)))
## n_na_macroregion
## 1 0
# W5: already has macroregion
wvs5 <- wvs5 %>%
mutate(macroregion_h = macroregion)
# W6: fold Pakistan into South Asia for harmonized displays
wvs6 <- wvs6 %>%
mutate(macroregion_h = dplyr::recode(region, "Pakistan" = "South Asia", .default = region))
# W7: already has macroregion (includes South Asia)
wvs7 <- wvs7 %>%
mutate(macroregion_h = macroregion)
# Quick verify
wvs5 %>% count(macroregion_h, sort = TRUE)
## # A tibble: 7 × 2
## macroregion_h n
## <chr> <int>
## 1 Western Europe Offshoots 20499
## 2 East Southeast Asia 13011
## 3 Latin America Caribbean 12589
## 4 Middle East North Africa MENA 12165
## 5 Sub-Saharan Africa 12097
## 6 Eastern Europe Post-Soviet 11613
## 7 India 2001
wvs6 %>% count(macroregion_h, sort = TRUE) # should include South Asia, not Pakistan
## macroregion_h n
## 1 Eastern Europe Post-Soviet 18410
## 2 Middle East North Africa MENA 16827
## 3 East Southeast Asia 13853
## 4 Latin America Caribbean 13435
## 5 Western Europe Offshoots 11893
## 6 Sub-Saharan Africa 9869
## 7 India 4078
## 8 South Asia 1200
wvs7 %>% count(macroregion_h, sort = TRUE)
## macroregion_h n
## 1 East Southeast Asia 23218
## 2 Western Europe Offshoots 19417
## 3 Latin America Caribbean 17439
## 4 Eastern Europe Post-Soviet 13951
## 5 Middle East North Africa MENA 12321
## 6 Sub-Saharan Africa 4948
## 7 South Asia 4234
## 8 India 1692
Economic ranking
clean_0_to_na <- function(x) {
x <- suppressWarnings(as.numeric(x))
ifelse(x == 0, NA, x)
}
compute_country_means <- function(df, country_var, weight_var, vars) {
df2 <- df %>%
mutate(across(all_of(vars), clean_0_to_na)) %>%
filter(!is.na(.data[[country_var]]), !is.na(.data[[weight_var]]))
des <- svydesign(ids = ~1, weights = as.formula(paste0("~", weight_var)), data = df2)
purrr::map_dfr(vars, function(v) {
out <- svyby(
formula = as.formula(paste0("~", v)),
by = as.formula(paste0("~", country_var)),
design = des,
FUN = svymean,
na.rm = TRUE
) %>%
as_tibble() %>%
rename(country = all_of(country_var)) %>%
transmute(variable = v, country, mean = .data[[v]])
out
})
}
compute_region_means_from_countrymeans <- function(country_means, country_to_region_df, region_var) {
country_means %>%
left_join(country_to_region_df, by = "country") %>%
filter(!is.na(.data[[region_var]])) %>%
group_by(.data[[region_var]], variable) %>%
summarise(
region_mean = mean(mean, na.rm = TRUE),
n_countries = n_distinct(country),
.groups = "drop"
) %>%
rename(region = all_of(region_var))
}
compute_india_stats <- function(country_means, india_code = 356) {
country_means %>%
group_by(variable) %>%
mutate(
n_countries = n_distinct(country),
rank_desc = dense_rank(desc(mean))
) %>%
ungroup() %>%
filter(country == india_code) %>%
transmute(variable, india_mean = mean, n_countries, india_rank = rank_desc)
}
# 0) Helpers (if not already defined in your cross-wave Rmd)
clean_0_to_na <- function(x) {
x <- suppressWarnings(as.numeric(x))
ifelse(x == 0, NA, x)
}
compute_country_means <- function(df, country_var, weight_var, vars) {
df2 <- df %>%
mutate(across(all_of(vars), clean_0_to_na)) %>%
filter(!is.na(.data[[country_var]]), !is.na(.data[[weight_var]]))
des <- survey::svydesign(ids = ~1, weights = as.formula(paste0("~", weight_var)), data = df2)
purrr::map_dfr(vars, function(v) {
survey::svyby(
formula = as.formula(paste0("~", v)),
by = as.formula(paste0("~", country_var)),
design = des,
FUN = survey::svymean,
na.rm = TRUE
) %>%
as_tibble() %>%
rename(country = all_of(country_var)) %>%
transmute(variable = v, country, mean = .data[[v]])
})
}
compute_region_means_from_countrymeans <- function(country_means, country_to_region_df, region_var) {
country_means %>%
left_join(country_to_region_df, by = "country") %>%
filter(!is.na(.data[[region_var]])) %>%
group_by(.data[[region_var]], variable) %>%
summarise(
region_mean = mean(mean, na.rm = TRUE),
n_countries = n_distinct(country),
.groups = "drop"
) %>%
rename(region = all_of(region_var))
}
compute_india_stats <- function(country_means, india_code = 356) {
country_means %>%
group_by(variable) %>%
mutate(
n_countries = n_distinct(country),
india_rank = dense_rank(desc(mean))
) %>%
ungroup() %>%
filter(country == india_code) %>%
transmute(variable, india_mean = mean, n_countries, india_rank)
}
compute_wave_outputs <- function(df, wave, year, country_var, weight_var, region_var, vars) {
df_work <- df %>%
filter(!is.na(.data[[region_var]]), !is.na(.data[[weight_var]]))
country_means <- compute_country_means(
df = df_work,
country_var = country_var,
weight_var = weight_var,
vars = vars
)
country_to_region <- df_work %>%
distinct(country = .data[[country_var]], region = .data[[region_var]]) %>%
rename(!!region_var := region)
region_means <- compute_region_means_from_countrymeans(
country_means = country_means,
country_to_region_df = country_to_region,
region_var = region_var
)
india_stats <- compute_india_stats(country_means, india_code = 356)
list(
country_means = country_means %>% mutate(wave = wave, year = year),
region_means = region_means %>% mutate(wave = wave, year = year),
india_stats = india_stats %>% mutate(wave = wave, year = year)
)
}
# 1) Economic item lists (match your wave files)
vars_w5 <- paste0("V", 116:121) # W5 economic system [file:1]
vars_w6 <- paste0("V", 96:101) # W6 economic system (NOT V95) [file:6]
vars_w7 <- paste0("Q", 106:111) # W7 economic system [file:2]
# 2) Run per-wave computations (using harmonized region var macroregion_h)
out_w5 <- compute_wave_outputs(wvs5, "W5", 2006, "V2", "V259", "macroregion_h", vars_w5)
out_w6 <- compute_wave_outputs(wvs6, "W6", 2012, "V2", "V258", "macroregion_h", vars_w6)
out_w7 <- compute_wave_outputs(wvs7, "W7", 2022, "B_COUNTRY","W_WEIGHT", "macroregion_h", vars_w7)
# 3) Combine India outputs into one table (rank is within-wave, among countries)
india_table <- bind_rows(out_w5$india_stats, out_w6$india_stats, out_w7$india_stats) %>%
arrange(variable, year)
india_table
## # A tibble: 18 × 6
## variable india_mean n_countries india_rank wave year
## <chr> <dbl> <int> <int> <chr> <dbl>
## 1 Q106 6.94 66 15 W7 2022
## 2 Q107 6.23 66 8 W7 2022
## 3 Q108 5.22 66 24 W7 2022
## 4 Q109 4.82 66 6 W7 2022
## 5 Q110 4.44 66 31 W7 2022
## 6 Q111 1.18 66 60 W7 2022
## 7 V100 3.30 60 52 W6 2012
## 8 V101 5.05 60 59 W6 2012
## 9 V116 3.75 58 57 W5 2006
## 10 V117 3.90 58 50 W5 2006
## 11 V118 3.50 58 52 W5 2006
## 12 V119 2.01 58 57 W5 2006
## 13 V120 2.88 58 52 W5 2006
## 14 V121 3.68 58 55 W5 2006
## 15 V96 2.92 60 60 W6 2012
## 16 V97 5.29 60 41 W6 2012
## 17 V98 3.36 60 52 W6 2012
## 18 V99 2.73 60 57 W6 2012
# 4) Combine region means into one table
region_table <- bind_rows(out_w5$region_means, out_w6$region_means, out_w7$region_means) %>%
arrange(year, variable, region)
library(dplyr)
library(tidyr)
library(stringr)
# Uses your existing helper
# clean_0_to_na <- function(x) { x <- suppressWarnings(as.numeric(x)); ifelse(x == 0, NA, x) }
# 1) Harmonised question key (concept-level; unmatched stay separate)
harm_key <- tibble::tribble(
~wave, ~variable, ~Questions,
"W5", "V116", "Income equality vs inequality incentives",
"W6", "V96", "Income equality vs inequality incentives",
"W7", "Q106", "Income equality vs inequality incentives",
"W5", "V117", "Private vs government ownership of business",
"W6", "V97", "Private vs government ownership of business",
"W7", "Q107", "Private vs government ownership of business",
"W5", "V118", "Government vs individual responsibility",
"W6", "V98", "Government vs individual responsibility",
"W7", "Q108", "Government vs individual responsibility",
"W5", "V119", "Competition good vs harmful",
"W6", "V99", "Competition good vs harmful",
"W7", "Q109", "Competition good vs harmful",
"W5", "V120", "Success hard work vs luck",
"W6", "V100", "Success hard work vs luck",
"W7", "Q110", "Success hard work vs luck",
"W5", "V121", "Wealth accumulation (others expense vs grows)",
"W6", "V101", "Wealth accumulation (others expense vs grows)",
"W7", "Q111", "Environment vs economic growth" # unmatched (Wave 7 only)
)
# 2) India valid N per question (unweighted N; after recoding 0->NA; requires non-missing weight too)
get_india_n_by_item <- function(df, wave, year, country_col, weight_col, vars) {
df %>%
filter(.data[[country_col]] == 356) %>%
mutate(across(all_of(vars), clean_0_to_na)) %>%
pivot_longer(cols = all_of(vars), names_to = "variable", values_to = "value") %>%
summarise(
india_n = sum(!is.na(value) & !is.na(.data[[weight_col]])),
.by = variable
) %>%
mutate(wave = wave, year = year)
}
india_n_w5 <- get_india_n_by_item(wvs5, "W5", 2006, "V2", "V259", vars_w5)
india_n_w6 <- get_india_n_by_item(wvs6, "W6", 2012, "V2", "V258", vars_w6)
india_n_w7 <- get_india_n_by_item(wvs7, "W7", 2022, "B_COUNTRY", "W_WEIGHT", vars_w7)
india_n_all <- bind_rows(india_n_w5, india_n_w6, india_n_w7)
# 3) India mean + India rank AMONG MACROREGIONS (computed from region_table)
india_macroregion_stats <- region_table %>%
group_by(wave, year, variable) %>%
mutate(
n_macroregions = n_distinct(region),
india_rank = dense_rank(desc(region_mean)) # 1 = highest mean among macroregions
) %>%
ungroup() %>%
filter(region == "India") %>%
transmute(wave, year, variable,
india_mean = round(region_mean, 2),
india_rank,
n_macroregions)
# 4) Build final long -> wide
india_final_table <- india_macroregion_stats %>%
left_join(india_n_all, by = c("wave","year","variable")) %>%
left_join(harm_key, by = c("wave","variable")) %>%
mutate(Questions = coalesce(Questions, variable)) %>%
select(Questions, year, india_rank, india_mean, india_n) %>%
mutate(year = as.character(year)) %>%
pivot_wider(
names_from = year,
values_from = c(india_rank, india_mean, india_n),
names_glue = "{year}_{.value}"
) %>%
select(
Questions,
`2006_india_rank`, `2012_india_rank`, `2022_india_rank`,
`2006_india_mean`, `2012_india_mean`, `2022_india_mean`,
`2006_india_n`, `2012_india_n`, `2022_india_n`
) %>%
rename(
`2006_rank` = `2006_india_rank`,
`2012_rank` = `2012_india_rank`,
`2022_rank` = `2022_india_rank`,
`2006_mean` = `2006_india_mean`,
`2012_mean` = `2012_india_mean`,
`2022_mean` = `2022_india_mean`,
`2006_n` = `2006_india_n`,
`2012_n` = `2012_india_n`,
`2022_n` = `2022_india_n`
) %>%
arrange(Questions)
india_final_table %>%
kable(
format = "html",
digits = 2,
caption = "India across waves: mean and rank among macroregions",
align = "lccccccccc"
) %>%
kable_styling(
full_width = FALSE,
bootstrap_options = c("striped", "hover", "condensed", "responsive")
) %>%
scroll_box(height = "350px", width = "100%")
India across waves: mean and rank among macroregions
|
Questions
|
2006_rank
|
2012_rank
|
2022_rank
|
2006_mean
|
2012_mean
|
2022_mean
|
2006_n
|
2012_n
|
2022_n
|
|
Competition good vs harmful
|
6
|
8
|
1
|
2.01
|
2.73
|
4.82
|
2001
|
3682
|
1692
|
|
Environment vs economic growth
|
NA
|
NA
|
8
|
NA
|
NA
|
1.18
|
NA
|
NA
|
1692
|
|
Government vs individual responsibility
|
6
|
8
|
3
|
3.50
|
3.36
|
5.22
|
2001
|
3723
|
1692
|
|
Income equality vs inequality incentives
|
7
|
8
|
2
|
3.75
|
2.92
|
6.94
|
2001
|
3780
|
1692
|
|
Private vs government ownership of business
|
6
|
6
|
1
|
3.90
|
5.29
|
6.23
|
2001
|
3583
|
1692
|
|
Success hard work vs luck
|
6
|
8
|
3
|
2.88
|
3.30
|
4.44
|
2001
|
3828
|
1692
|
|
Wealth accumulation (others expense vs grows)
|
6
|
8
|
NA
|
3.68
|
5.05
|
NA
|
2001
|
3419
|
NA
|
Democratic system
# Assumes already in memory:
# wvs5, wvs6, wvs7
# macroregion_h exists in each wave data
# Wave 7 uses B_COUNTRY and W_WEIGHT (your chosen convention)
# ----------------------------
# 0) Item lists (Democratic system)
# ----------------------------
vars_w5_dem <- paste0("V", 152:162) # W5 democratic system block[1]
vars_w6_dem <- paste0("V", 131:140) # W6 democratic system block[2]
vars_w7_dem <- paste0("Q", 241:250) # W7 democratic system items (excluding Q240 left-right)[3]
# Guardrails (fail loudly if names not present)
stopifnot(all(vars_w5_dem %in% names(wvs5)))
stopifnot(all(vars_w6_dem %in% names(wvs6)))
stopifnot(all(vars_w7_dem %in% names(wvs7)))
stopifnot(all(c("B_COUNTRY","W_WEIGHT","macroregion_h") %in% names(wvs7)))
# ----------------------------
# 1) Cleaning rule for demo variables
# W5 uses 0 as invalid; W6/W7 also have negative codes in some items.
# We'll treat <=0 as missing for all waves.
# ----------------------------
clean_demo_to_na <- function(x) {
x <- suppressWarnings(as.numeric(x))
ifelse(x <= 0, NA, x)
}
# ----------------------------
# 2) Country means (survey-weighted)
# ----------------------------
compute_country_means <- function(df, country_var, weight_var, vars) {
df2 <- df %>%
mutate(across(all_of(vars), clean_demo_to_na)) %>%
filter(!is.na(.data[[country_var]]), !is.na(.data[[weight_var]]))
des <- svydesign(ids = ~1, weights = as.formula(paste0("~", weight_var)), data = df2)
purrr::map_dfr(vars, function(v) {
svyby(
formula = as.formula(paste0("~", v)),
by = as.formula(paste0("~", country_var)),
design = des,
FUN = svymean,
na.rm = TRUE
) %>%
as_tibble() %>%
rename(country = all_of(country_var)) %>%
transmute(variable = v, country, mean = .data[[v]])
})
}
# ----------------------------
# 3) Macroregion mean = average of country means
# ----------------------------
compute_region_means_from_countrymeans <- function(country_means, country_to_region_df, region_var) {
country_means %>%
left_join(country_to_region_df, by = "country") %>%
filter(!is.na(.data[[region_var]])) %>%
group_by(.data[[region_var]], variable) %>%
summarise(
region_mean = mean(mean, na.rm = TRUE),
n_countries = n_distinct(country),
.groups = "drop"
) %>%
rename(region = all_of(region_var))
}
compute_wave_region_table <- function(df, wave, year, country_var, weight_var, region_var, vars) {
df_work <- df %>%
filter(!is.na(.data[[region_var]]), !is.na(.data[[weight_var]]))
country_means <- compute_country_means(
df = df_work,
country_var = country_var,
weight_var = weight_var,
vars = vars
)
country_to_region <- df_work %>%
distinct(country = .data[[country_var]], region = .data[[region_var]]) %>%
rename(!!region_var := region)
compute_region_means_from_countrymeans(
country_means = country_means,
country_to_region_df = country_to_region,
region_var = region_var
) %>%
mutate(wave = wave, year = year)
}
region_table_demo <- bind_rows(
compute_wave_region_table(wvs5, "W5", 2006, "V2", "V259", "macroregion_h", vars_w5_dem),
compute_wave_region_table(wvs6, "W6", 2012, "V2", "V258", "macroregion_h", vars_w6_dem),
compute_wave_region_table(wvs7, "W7", 2022, "B_COUNTRY", "W_WEIGHT","macroregion_h", vars_w7_dem)
) %>%
arrange(year, variable, region)
# ----------------------------
# 4) Harmonised label key (concept-level)
# (Shared across waves; W5 has a few extras, kept as W5-only rows.)
# ----------------------------
harm_key_demo <- tibble::tribble(
~wave, ~variable, ~Questions,
"W5","V152","Democracy: Tax rich subsidize poor (essential)",
"W6","V131","Democracy: Tax rich subsidize poor (essential)",
"W7","Q241","Democracy: Tax rich subsidize poor (essential)",
"W5","V153","Democracy: Religious authorities interpret laws (essential)",
"W6","V132","Democracy: Religious authorities interpret laws (essential)",
"W7","Q242","Democracy: Religious authorities interpret laws (essential)",
"W5","V154","Democracy: Free elections for leaders (essential)",
"W6","V133","Democracy: Free elections for leaders (essential)",
"W7","Q243","Democracy: Free elections for leaders (essential)",
"W5","V155","Democracy: State aid for unemployment (essential)",
"W6","V134","Democracy: State aid for unemployment (essential)",
"W7","Q244","Democracy: State aid for unemployment (essential)",
"W5","V156","Democracy: Army takeover when govt incompetent (essential)",
"W6","V135","Democracy: Army takeover when govt incompetent (essential)",
"W7","Q245","Democracy: Army takeover when govt incompetent (essential)",
"W5","V157","Democracy: Civil rights protect against oppression (essential)",
"W6","V136","Democracy: Civil rights protect against oppression (essential)",
"W7","Q246","Democracy: Civil rights protect against oppression (essential)",
"W5","V158","Democracy: Economy is prospering (essential)", # W5-only[1]
"W5","V159","Democracy: Criminals are severely punished (essential)", # W5-only[1]
"W5","V160","Democracy: Referendums can change laws (essential)", # W5-only[1]
"W5","V161","Democracy: Women have same rights as men (essential)",
"W6","V139","Democracy: Women equal rights (essential)",
"W7","Q249","Democracy: Women equal rights (essential)",
"W5","V162","Importance of democracy",
"W6","V140","Importance of democracy",
"W7","Q250","Importance of democratic country"
)
# ----------------------------
# 5) India valid N per item (unweighted N; requires non-missing weight; <=0 -> NA)
# ----------------------------
get_india_n_by_item <- function(df, wave, year, country_col, weight_col, vars) {
df %>%
filter(.data[[country_col]] == 356) %>%
mutate(across(all_of(vars), clean_demo_to_na)) %>%
pivot_longer(cols = all_of(vars), names_to = "variable", values_to = "value") %>%
summarise(india_n = sum(!is.na(value) & !is.na(.data[[weight_col]])), .by = variable) %>%
mutate(wave = wave, year = year)
}
india_n_demo <- bind_rows(
get_india_n_by_item(wvs5, "W5", 2006, "V2", "V259", vars_w5_dem),
get_india_n_by_item(wvs6, "W6", 2012, "V2", "V258", vars_w6_dem),
get_india_n_by_item(wvs7, "W7", 2022, "B_COUNTRY", "W_WEIGHT", vars_w7_dem)
)
# ----------------------------
# 6) India mean + India rank among macroregions (within wave×variable)
# ----------------------------
india_macro_stats_demo <- region_table_demo %>%
group_by(wave, year, variable) %>%
mutate(
n_macroregions = n_distinct(region),
india_rank = dense_rank(desc(region_mean)) # 1 = highest mean among macroregions
) %>%
ungroup() %>%
filter(region == "India") %>%
transmute(
wave, year, variable,
india_mean = round(region_mean, 2),
india_rank,
n_macroregions
)
# ----------------------------
# 7) Final long -> wide table + scrollable kable
# ----------------------------
india_demo_table <- india_macro_stats_demo %>%
left_join(india_n_demo, by = c("wave","year","variable")) %>%
left_join(harm_key_demo, by = c("wave","variable")) %>%
mutate(Questions = coalesce(Questions, variable)) %>%
select(Questions, year, india_rank, india_mean, india_n) %>%
mutate(year = as.character(year)) %>%
pivot_wider(
names_from = year,
values_from = c(india_rank, india_mean, india_n),
names_glue = "{year}_{.value}"
) %>%
transmute(
Questions,
`2006_rank` = `2006_india_rank`,
`2012_rank` = `2012_india_rank`,
`2022_rank` = `2022_india_rank`,
`2006_mean` = `2006_india_mean`,
`2012_mean` = `2012_india_mean`,
`2022_mean` = `2022_india_mean`,
`2006_n` = `2006_india_n`,
`2012_n` = `2012_india_n`,
`2022_n` = `2022_india_n`
) %>%
arrange(Questions)
india_demo_table %>%
kable(
format = "html",
digits = 2,
caption = "India across waves: Democratic system (mean + rank among macroregions)"
) %>%
kable_styling(
full_width = FALSE,
bootstrap_options = c("striped", "hover", "condensed", "responsive")
) %>%
scroll_box(height = "450px", width = "100%")
India across waves: Democratic system (mean + rank among macroregions)
|
Questions
|
2006_rank
|
2012_rank
|
2022_rank
|
2006_mean
|
2012_mean
|
2022_mean
|
2006_n
|
2012_n
|
2022_n
|
|
Democracy: Army takeover when govt incompetent (essential)
|
2
|
2
|
3
|
5.47
|
5.70
|
5.49
|
1362
|
4005
|
1309
|
|
Democracy: Civil rights protect against oppression (essential)
|
3
|
6
|
5
|
7.95
|
7.27
|
7.54
|
1420
|
4004
|
1476
|
|
Democracy: Criminals are severely punished (essential)
|
4
|
NA
|
NA
|
7.53
|
NA
|
NA
|
1578
|
NA
|
NA
|
|
Democracy: Economy is prospering (essential)
|
5
|
NA
|
NA
|
7.39
|
NA
|
NA
|
1441
|
NA
|
NA
|
|
Democracy: Free elections for leaders (essential)
|
2
|
6
|
3
|
8.62
|
7.82
|
8.12
|
1697
|
4009
|
1581
|
|
Democracy: Referendums can change laws (essential)
|
4
|
NA
|
NA
|
7.36
|
NA
|
NA
|
1446
|
NA
|
NA
|
|
Democracy: Religious authorities interpret laws (essential)
|
4
|
5
|
4
|
4.17
|
4.00
|
4.69
|
1436
|
4003
|
1203
|
|
Democracy: State aid for unemployment (essential)
|
1
|
3
|
5
|
8.30
|
7.37
|
7.19
|
1663
|
4001
|
1569
|
|
Democracy: Tax rich subsidize poor (essential)
|
1
|
2
|
2
|
8.13
|
7.05
|
7.35
|
1687
|
4014
|
1512
|
|
Democracy: Women equal rights (essential)
|
NA
|
3
|
3
|
NA
|
8.11
|
8.18
|
NA
|
4012
|
1601
|
|
Democracy: Women have same rights as men (essential)
|
2
|
NA
|
NA
|
8.21
|
NA
|
NA
|
1620
|
NA
|
NA
|
|
Importance of democracy
|
7
|
8
|
NA
|
7.08
|
7.77
|
NA
|
1609
|
4045
|
NA
|
|
Importance of democratic country
|
NA
|
NA
|
3
|
NA
|
NA
|
8.43
|
NA
|
NA
|
1612
|
|
Q247
|
NA
|
NA
|
3
|
NA
|
NA
|
6.51
|
NA
|
NA
|
1488
|
|
Q248
|
NA
|
NA
|
3
|
NA
|
NA
|
6.65
|
NA
|
NA
|
1519
|
|
V137
|
NA
|
1
|
NA
|
NA
|
7.10
|
NA
|
NA
|
3993
|
NA
|
|
V138
|
NA
|
4
|
NA
|
NA
|
6.41
|
NA
|
NA
|
3996
|
NA
|
Neighbour
calc_india_props <- function(dat, year, country_var, weight_var, items_named_vec) {
vars <- names(items_named_vec)
# Clean only necessary variables
dat_i <- dat %>%
mutate(
across(all_of(c(country_var, weight_var, vars)),
~ suppressWarnings(as.numeric(zap_labels(.x))))
) %>%
filter(.data[[country_var]] == 356,
!is.na(.data[[weight_var]]))
# Survey design
des <- svydesign(
ids = ~1,
weights = as.formula(paste0("~", weight_var)),
data = dat_i
)
# Estimate proportion where response == 1 (REJECTION)
purrr::imap_dfr(items_named_vec, function(label, v) {
est <- svymean(
as.formula(paste0("~I(", v, " == 1)")),
des,
na.rm = TRUE
)
rejection_rate <- as.numeric(coef(est)[1]) * 100
tibble(
year = as.character(year),
item = label,
value = round(rejection_rate) # nearest integer
)
})
}
# ---------------------------
# CORRECT VARIABLE MAPPING
# ---------------------------
items_w5 <- c(
V34 = "Drug addicts",
V35 = "People of a different race",
V36 = "People who have AIDS",
V37 = "Immigrants/foreign workers",
V38 = "Homosexuals",
V39 = "People of a different religion"
)
items_w6 <- c(
V36 = "Drug addicts",
V37 = "People of a different race",
V38 = "People who have AIDS",
V39 = "Immigrants/foreign workers",
V40 = "Homosexuals",
V41 = "People of a different religion"
)
items_w7 <- c(
Q27 = "Drug addicts",
Q28 = "People of a different race",
Q29 = "People who have AIDS",
Q30 = "Immigrants/foreign workers",
Q31 = "Homosexuals",
Q32 = "People of a different religion"
)
# ---------------------------
# Resolve W7 column names
# ---------------------------
country_var_w7 <- if ("B_COUNTRY" %in% names(wvs7)) "B_COUNTRY" else "BCOUNTRY"
weight_var_w7 <- if ("W_WEIGHT" %in% names(wvs7)) "W_WEIGHT" else "WWEIGHT"
# ---------------------------
# Combine waves
# ---------------------------
res_long <- bind_rows(
calc_india_props(wvs5, 2006, "V2", "V259", items_w5),
calc_india_props(wvs6, 2012, "V2", "V258", items_w6),
calc_india_props(wvs7, 2022, country_var_w7, weight_var_w7, items_w7)
)
# ---------------------------
# FINAL TABLE FORMAT
# ---------------------------
res_wide <- res_long %>%
pivot_wider(names_from = item, values_from = value) %>%
arrange(year)
res_wide %>%
knitr::kable(
format = "html",
digits = 1,
caption = "India: % mentioning each group as undesirable neighbour (survey-weighted)"
) %>%
kableExtra::kable_styling(
full_width = FALSE,
bootstrap_options = c("striped","hover","condensed","responsive")
)
India: % mentioning each group as undesirable neighbour
(survey-weighted)
|
year
|
Drug addicts
|
People of a different race
|
People who have AIDS
|
Immigrants/foreign workers
|
Homosexuals
|
People of a different religion
|
|
2006
|
46
|
56
|
56
|
65
|
60
|
56
|
|
2012
|
6
|
74
|
38
|
53
|
35
|
72
|
|
2022
|
22
|
66
|
77
|
83
|
79
|
75
|
Moral permisiveness
moral_w5_named <- c(
V198 = "Government benefits",
V199 = "Public transport fare",
V200 = "Tax evasion",
V201 = "Bribing",
V202 = "Homosexuality",
V203 = "Prostitution",
V204 = "Abortion",
V205 = "Divorce",
V206 = "Euthanasia",
V207 = "Suicide",
V208 = "Domestic violence"
)
moral_w6_named <- c(
V198 = "Government benefits",
V199 = "Public transport fare",
V201 = "Tax evasion",
V202 = "Bribing",
V203 = "Homosexuality",
V203A = "Prostitution",
V204 = "Abortion",
V205 = "Divorce",
V206 = "Premarital sex",
V207 = "Suicide",
V207A = "Euthanasia",
V208 = "Domestic violence"
)
moral_w7_named <- c(
Q177 = "Government benefits",
Q178 = "Public transport fare",
Q180 = "Tax evasion",
Q181 = "Bribing",
Q182 = "Homosexuality",
Q183 = "Prostitution",
Q184 = "Abortion",
Q185 = "Divorce",
Q186 = "Premarital sex",
Q187 = "Suicide",
Q188 = "Euthanasia",
Q189 = "Domestic violence"
)
# ── Sanity check ───────────────────────────────────────────────────────────────
stopifnot(all(names(moral_w5_named) %in% names(wvs5)))
stopifnot(all(names(moral_w6_named) %in% names(wvs6)))
stopifnot(all(names(moral_w7_named) %in% names(wvs7)))
# ── Survey-weighted India means ────────────────────────────────────────────────
calc_india_moral <- function(dat, year, country_var, weight_var, named_vars) {
vars <- names(named_vars)
dat_i <- dat %>%
mutate(across(
all_of(vars),
~ { x <- suppressWarnings(as.numeric(zap_labels(.x))); ifelse(x <= 0, NA, x) }
)) %>%
filter(.data[[country_var]] == 356,
!is.na(.data[[weight_var]]))
des <- svydesign(
ids = ~1,
weights = as.formula(paste0("~", weight_var)),
data = dat_i
)
imap_dfr(named_vars, function(label, v) {
if (sum(!is.na(dat_i[[v]])) == 0)
return(tibble(year = as.character(year), item = label, value = NA_real_))
est <- svymean(as.formula(paste0("~", v)), des, na.rm = TRUE)
tibble(
year = as.character(year),
item = label,
value = round(as.numeric(coef(est)[1]), 1)
)
})
}
# ── Run all waves ──────────────────────────────────────────────────────────────
moral_long <- bind_rows(
calc_india_moral(wvs5, 2006, "V2", "V259", moral_w5_named),
calc_india_moral(wvs6, 2012, "V2", "V258", moral_w6_named),
calc_india_moral(wvs7, 2022, "B_COUNTRY", "W_WEIGHT", moral_w7_named)
)
# ── Wide: Year as rows, items as columns ───────────────────────────────────────
item_order <- c(
"Government benefits",
"Public transport fare",
"Tax evasion",
"Bribing",
"Homosexuality",
"Prostitution",
"Abortion",
"Divorce",
"Premartial sex",
"Suicide",
"Euthanasia",
"Domestic voilence"
)
moral_table <- moral_long %>%
mutate(item = factor(item, levels = item_order)) %>%
pivot_wider(names_from = item, values_from = value) %>%
arrange(year) %>%
rename(Year = year)
## Warning: Values from `value` are not uniquely identified; output will contain list-cols.
## • Use `values_fn = list` to suppress this warning.
## • Use `values_fn = {summary_fun}` to summarise duplicates.
## • Use the following dplyr code to identify duplicates.
## {data} |>
## dplyr::summarise(n = dplyr::n(), .by = c(year, item)) |>
## dplyr::filter(n > 1L)
# ── Render ─────────────────────────────────────────────────────────────────────
moral_table %>%
kable(
format = "html",
digits = 1,
caption = "India: Moral Permissiveness (1 = Never justifiable, 10 = Always justifiable)"
) %>%
kable_styling(
full_width = FALSE,
bootstrap_options = c("striped", "hover", "condensed", "responsive")
) %>%
column_spec(1, bold = TRUE) %>%
scroll_box(height = "200px", width = "100%")
India: Moral Permissiveness (1 = Never justifiable, 10 = Always
justifiable)
|
Year
|
Government benefits
|
Public transport fare
|
Tax evasion
|
Bribing
|
Homosexuality
|
Prostitution
|
Abortion
|
Divorce
|
Euthanasia
|
Suicide
|
NA
|
|
2006
|
3.1
|
3.3
|
3
|
3
|
3
|
3.1
|
3.3
|
3.9
|
3.9
|
3.4
|
3.1
|
|
2012
|
2.5
|
1.7
|
1.7
|
1.6
|
1.6
|
1.6
|
1.6
|
1.9
|
NA
|
1.6
|
1.6, 1.8
|
|
2022
|
2.7
|
2.3
|
2
|
1.9
|
2.7
|
2.2
|
2.8
|
3.1
|
2.9
|
1.9
|
2.2, 2.0
|
Institutional trust
trust_w5_named <- c(
V131 = "Religious institutions",
V132 = "Armed forces",
V133 = "Press",
V134 = "Television",
V136 = "Police",
V137 = "Courts",
V138 = "Government",
V139 = "Political parties",
V140 = "Parliament",
V141 = "Civil service"
)
trust_w6_named <- c(
V108 = "Religious institutions",
V109 = "Armed forces",
V110 = "Press",
V111 = "Television",
V113 = "Police",
V114 = "Courts",
V115 = "Government",
V116 = "Political parties",
V117 = "Parliament",
V118 = "Civil service"
)
trust_w7_named <- c(
Q64 = "Religious institutions",
Q65 = "Armed forces",
Q66 = "Press",
Q67 = "Television",
Q69 = "Police",
Q70 = "Courts",
Q71 = "Government",
Q72 = "Political parties",
Q73 = "Parliament",
Q74 = "Civil service"
)
# ─────────────────────────────────────────────
# INSTITUTIONAL TRUST – HIGH CONFIDENCE SHARE
# (1 & 2 collapsed)
# ─────────────────────────────────────────────
# ── Sanity check ─────────────────────────────
stopifnot(all(names(trust_w5_named) %in% names(wvs5)))
stopifnot(all(names(trust_w6_named) %in% names(wvs6)))
stopifnot(all(names(trust_w7_named) %in% names(wvs7)))
calc_india_trust <- function(dat, year, country_var, weight_var, named_vars) {
vars <- names(named_vars)
dat_i <- dat %>%
mutate(across(
all_of(vars),
~ {
x <- suppressWarnings(as.numeric(zap_labels(.x)))
case_when(
x %in% c(1, 2) ~ 1, # High confidence
x %in% c(3, 4) ~ 0, # Low confidence
TRUE ~ NA_real_
)
}
)) %>%
filter(.data[[country_var]] == 356,
!is.na(.data[[weight_var]]))
des <- svydesign(
ids = ~1,
weights = as.formula(paste0("~", weight_var)),
data = dat_i
)
imap_dfr(named_vars, function(label, v) {
if (sum(!is.na(dat_i[[v]])) == 0)
return(tibble(year = as.character(year), item = label, value = NA_real_))
est <- svymean(as.formula(paste0("~", v)), des, na.rm = TRUE)
tibble(
year = as.character(year),
item = label,
value = round(as.numeric(coef(est)[1]) * 100, 1) # percentage
)
})
}
# ── Run all waves ────────────────────────────
trust_long <- bind_rows(
calc_india_trust(wvs5, 2006, "V2", "V259", trust_w5_named),
calc_india_trust(wvs6, 2012, "V2", "V258", trust_w6_named),
calc_india_trust(wvs7, 2022, "B_COUNTRY", "W_WEIGHT", trust_w7_named)
)
institution_order <- c(
"Parliament",
"Government",
"Political parties",
"Civil service",
"Armed forces",
"Police",
"Courts",
"Press",
"Television",
"Religious institutions"
)
trust_table <- trust_long %>%
mutate(item = factor(item, levels = institution_order)) %>%
pivot_wider(names_from = item, values_from = value) %>%
arrange(year) %>%
rename(Year = year)
trust_table %>%
kable(
format = "html",
digits = 1,
caption = "India: High Confidence in Institutions (% saying 'A great deal' or 'Quite a lot').
Harmonised across WVS 2006–2022."
) %>%
kable_styling(
full_width = FALSE,
bootstrap_options = c("striped", "hover", "condensed", "responsive")
) %>%
column_spec(1, bold = TRUE) %>%
scroll_box(height = "200px", width = "100%")
India: High Confidence in Institutions (% saying ‘A great deal’ or
‘Quite a lot’). Harmonised across WVS 2006–2022.
|
Year
|
Religious institutions
|
Armed forces
|
Press
|
Television
|
Police
|
Courts
|
Government
|
Political parties
|
Parliament
|
Civil service
|
|
2006
|
83.4
|
83.3
|
75.8
|
74.9
|
64.1
|
68.9
|
54.9
|
46.4
|
62.4
|
54.3
|
|
2012
|
96.1
|
87.1
|
71.8
|
74.5
|
51.2
|
64.5
|
50.4
|
37.4
|
58.4
|
61.0
|
|
2022
|
89.7
|
86.9
|
65.8
|
63.4
|
66.8
|
74.6
|
65.1
|
42.3
|
73.8
|
80.6
|
Regime preferences
clean_reg <- function(x) {
x <- suppressWarnings(as.numeric(haven::zap_labels(x)))
x[x < 0] <- NA; x
}
regime_cfg <- list(
list(year="2006", dat=quote(wvs5), ctry="V2", wt="V259",
vars=c(V148="Strong leader", V149="Experts", V150="Army rule", V151="Democratic system")),
list(year="2012", dat=quote(wvs6), ctry="V2", wt="V258",
vars=c(V127="Strong leader", V128="Experts", V129="Army rule", V130="Democratic system")),
list(year="2022", dat=quote(wvs7), ctry="B_COUNTRY", wt="W_WEIGHT",
vars=c(Q235="Strong leader", Q236="Experts", Q237="Army rule", Q238="Democratic system"))
)
compute_block <- function(cfg) {
dat <- eval(cfg$dat)
vars <- names(cfg$vars)
labs <- cfg$vars
dat_c <- dat %>%
mutate(
across(all_of(c(cfg$ctry, cfg$wt)), ~suppressWarnings(as.numeric(haven::zap_labels(.x)))),
across(all_of(vars), clean_reg)
) %>%
mutate(across(all_of(vars),
~as.integer(.x %in% c(1, 2)),
.names="acc_{.col}")) %>%
filter(!is.na(.data[[cfg$wt]]))
# ── India % acceptable ──────────────────────────────────────────────
india_dat <- dat_c %>% filter(.data[[cfg$ctry]] == 356)
des_i <- svydesign(ids=~1, weights=as.formula(paste0("~", cfg$wt)), data=india_dat)
india_pct <- imap_dfr(labs, function(lbl, v) {
bv <- paste0("acc_", v)
est <- svymean(as.formula(paste0("~", bv)), des_i, na.rm=TRUE)
tibble(question=lbl, india_pct=round(as.numeric(coef(est)) * 100, 1))
})
# ── Region country-weighted averages ───────────────────────────────
rgn_dat <- dat_c %>% filter(!is.na(macroregion_h), macroregion_h != "India")
region_avgs <- imap_dfr(labs, function(lbl, v) {
bv <- paste0("acc_", v)
d <- rgn_dat %>% filter(!is.na(.data[[bv]]))
des <- svydesign(ids=~1, weights=as.formula(paste0("~", cfg$wt)), data=d)
svyby(as.formula(paste0("~", bv)),
as.formula(paste0("~", cfg$ctry, " + macroregion_h")),
des, svymean, na.rm=TRUE) %>%
as.data.frame() %>%
select(macroregion_h, cpct=all_of(bv)) %>%
mutate(label=lbl)
}) %>%
group_by(macroregion_h, label) %>%
summarise(rpct=mean(cpct, na.rm=TRUE) * 100, .groups="drop")
n_entities <- n_distinct(region_avgs$macroregion_h) + 1
# ── Rank India ──────────────────────────────────────────────────────
ranking <- region_avgs %>%
bind_rows(india_pct %>% transmute(macroregion_h="India", label=question, rpct=india_pct)) %>%
group_by(label) %>%
mutate(rank=rank(-rpct, ties.method="min")) %>%
filter(macroregion_h=="India") %>%
ungroup() %>%
select(question=label, rank) %>%
mutate(n_entities=n_entities)
india_pct %>%
left_join(ranking, by="question") %>%
mutate(year=cfg$year)
}
results <- map_dfr(regime_cfg, compute_block)
stopifnot(nrow(results %>% count(question, year) %>% filter(n > 1)) == 0)
res_wide <- results %>%
mutate(rank_fmt=paste0(rank, "/", n_entities)) %>%
select(question, year, rank_fmt, india_pct) %>%
pivot_wider(
names_from = year,
values_from = c(rank_fmt, india_pct),
names_glue = "{year}_{.value}"
) %>%
select(question,
`2006_rank_fmt`, `2012_rank_fmt`, `2022_rank_fmt`,
`2006_india_pct`, `2012_india_pct`, `2022_india_pct`) %>%
mutate(question=factor(question,
levels=c("Strong leader","Experts","Army rule","Democratic system"))) %>%
arrange(question)
res_wide %>%
knitr::kable(
format = "html",
col.names = c("Regime type",
"2006 Rank","2012 Rank","2022 Rank",
"2006%","2012%","2022%"),
caption = "India: % finding each regime 'acceptable' (Very/Fairly good) & rank among macro-regions"
) %>%
kableExtra::kable_styling(
full_width = FALSE,
bootstrap_options = c("striped","hover","condensed")
) %>%
kableExtra::add_header_above(
c(" "=1, "Rank among macro-regions"=3, "% Acceptable — India"=3)
) %>%
kableExtra::column_spec(1, bold=TRUE)
India: % finding each regime ‘acceptable’ (Very/Fairly good) & rank
among macro-regions
|
|
Rank among macro-regions
|
% Acceptable — India
|
|
Regime type
|
2006 Rank
|
2012 Rank
|
2022 Rank
|
2006%
|
2012%
|
2022%
|
|
Strong leader
|
1/7
|
1/8
|
1/8
|
45.6
|
56.4
|
62.9
|
|
Experts
|
4/7
|
2/8
|
2/8
|
50.0
|
59.4
|
63.2
|
|
Army rule
|
3/7
|
2/8
|
4/8
|
24.8
|
37.8
|
33.7
|
|
Democratic system
|
7/7
|
5/8
|
6/8
|
70.0
|
79.7
|
77.2
|
Jobs scarce
clean_reg <- function(x) {
x <- suppressWarnings(as.numeric(haven::zap_labels(x)))
x[x < 0] <- NA; x
}
jobs_cfg <- list(
# W5: 3-pt (1=Agree, 2=Neither, 3=Disagree)
list(year="2006", dat=quote(wvs5), ctry="V2", wt="V259",
agree_codes=c(1),
vars=c(V44="Men over women", V45="Nationals over immigrants")),
# W6: 3-pt (1=Agree, 2=Neither, 3=Disagree)
list(year="2012", dat=quote(wvs6), ctry="V2", wt="V258",
agree_codes=c(1),
vars=c(V43="Men over women", V44="Nationals over immigrants")),
# W7: 5-pt — collapse 1 (Strongly agree) + 2 (Agree) → "Agree"
list(year="2022", dat=quote(wvs7), ctry="B_COUNTRY", wt="W_WEIGHT",
agree_codes=c(1, 2),
vars=c(Q33="Men over women", Q34="Nationals over immigrants"))
)
compute_jobs_block <- function(cfg) {
dat <- eval(cfg$dat)
vars <- names(cfg$vars)
labs <- cfg$vars
ac <- cfg$agree_codes
dat_c <- dat %>%
mutate(
across(all_of(c(cfg$ctry, cfg$wt)), ~suppressWarnings(as.numeric(haven::zap_labels(.x)))),
across(all_of(vars), clean_reg)
) %>%
mutate(across(all_of(vars),
~as.integer(.x %in% ac),
.names="acc_{.col}")) %>%
filter(!is.na(.data[[cfg$wt]]))
# ── India % agree ───────────────────────────────────────────────────
india_dat <- dat_c %>% filter(.data[[cfg$ctry]] == 356)
des_i <- svydesign(ids=~1, weights=as.formula(paste0("~", cfg$wt)), data=india_dat)
india_pct <- imap_dfr(labs, function(lbl, v) {
bv <- paste0("acc_", v)
est <- svymean(as.formula(paste0("~", bv)), des_i, na.rm=TRUE)
tibble(question=lbl, india_pct=round(as.numeric(coef(est)) * 100, 1))
})
# ── Region country-weighted averages ───────────────────────────────
rgn_dat <- dat_c %>% filter(!is.na(macroregion_h), macroregion_h != "India")
region_avgs <- imap_dfr(labs, function(lbl, v) {
bv <- paste0("acc_", v)
d <- rgn_dat %>% filter(!is.na(.data[[bv]]))
des <- svydesign(ids=~1, weights=as.formula(paste0("~", cfg$wt)), data=d)
svyby(as.formula(paste0("~", bv)),
as.formula(paste0("~", cfg$ctry, " + macroregion_h")),
des, svymean, na.rm=TRUE) %>%
as.data.frame() %>%
select(macroregion_h, cpct=all_of(bv)) %>%
mutate(label=lbl)
}) %>%
group_by(macroregion_h, label) %>%
summarise(rpct=mean(cpct, na.rm=TRUE) * 100, .groups="drop")
n_entities <- n_distinct(region_avgs$macroregion_h) + 1
# ── Rank India ──────────────────────────────────────────────────────
ranking <- region_avgs %>%
bind_rows(india_pct %>% transmute(macroregion_h="India", label=question, rpct=india_pct)) %>%
group_by(label) %>%
mutate(rank=rank(-rpct, ties.method="min")) %>%
filter(macroregion_h=="India") %>%
ungroup() %>%
select(question=label, rank) %>%
mutate(n_entities=n_entities)
india_pct %>%
left_join(ranking, by="question") %>%
mutate(year=cfg$year)
}
results_jobs <- map_dfr(jobs_cfg, compute_jobs_block)
stopifnot(nrow(results_jobs %>% count(question, year) %>% filter(n > 1)) == 0)
res_wide_jobs <- results_jobs %>%
mutate(rank_fmt=paste0(rank, "/", n_entities)) %>%
select(question, year, rank_fmt, india_pct) %>%
pivot_wider(
names_from = year,
values_from = c(rank_fmt, india_pct),
names_glue = "{year}_{.value}"
) %>%
select(question,
`2006_rank_fmt`, `2012_rank_fmt`, `2022_rank_fmt`,
`2006_india_pct`, `2012_india_pct`, `2022_india_pct`) %>%
mutate(question=factor(question,
levels=c("Men over women", "Nationals over immigrants"))) %>%
arrange(question)
res_wide_jobs %>%
knitr::kable(
format = "html",
col.names = c("Statement",
"2006 Rank","2012 Rank","2022 Rank",
"2006%","2012%","2022%"),
caption = "India: % agreeing jobs should go to men/nationals when scarce & rank among macro-regions"
) %>%
kableExtra::kable_styling(
full_width = FALSE,
bootstrap_options = c("striped","hover","condensed")
) %>%
kableExtra::add_header_above(
c(" "=1, "Rank among macro-regions"=3, "% Agree — India"=3)
) %>%
kableExtra::column_spec(1, bold=TRUE) %>%
kableExtra::footnote(
general = paste0(
"2006 & 2012: Agree on 3-point scale (Agree / Neither / Disagree). ",
"2022: Strongly agree + Agree collapsed on 5-point scale."
),
general_title = "Note: "
)
India: % agreeing jobs should go to men/nationals when scarce & rank
among macro-regions
|
|
Rank among macro-regions
|
% Agree — India
|
|
Statement
|
2006 Rank
|
2012 Rank
|
2022 Rank
|
2006%
|
2012%
|
2022%
|
|
Men over women
|
2/7
|
1/8
|
3/8
|
50.5
|
72.7
|
56.8
|
|
Nationals over immigrants
|
1/7
|
1/8
|
6/8
|
74.2
|
32.7
|
71.6
|
|
Note:
|
|
2006 & 2012: Agree on 3-point scale (Agree / Neither /
Disagree). 2022: Strongly agree + Agree collapsed on 5-point scale.
|