Replication Codes for ‘Do democracies prefer to support invaded countries that are democratic?’
Harukata Takenaka
November 13, 2025
Replication Materials
This repository provides the data and code required to reproduce the analyses and figures reported in “Do democracies prefer to support invaded countries that are democratic?” All scripts are written in R and are organized to mirror the structure of the paper and Supplementary Information.
Required R Packages
We conducted the analyses in R (Version 4.4.2). The packages listed below are needed to reproduce the results. You can install any missing packages with the code snippet provided. We recommend using renv to snapshot and restore the exact package versions.
library(broom) #Ver.1.0.7
library(flextable) #Ver.0.9.7
library(interactions) #Ver.
library(lubridate) #Ver.1.2.0
library(moments) #Ver.0.14.1
library(officer) #Ver.0.6.7
library(patchwork) #Ver.1.3.0
library(psych) #Ver.2.4.12
library(scales) #Ver.1.4.0
library(stargazer) #Ver.5.2.3
library(tidyverse) #Ver.2.0.0
library(visreg) #Ver.2.7.0
library(vtable) #Ver.1.4.8Setting Variables
Setting variables for the Study 1
security <- read.csv("security.csv") %>%
filter(Q6.1_10 == 2) %>%
mutate(
across(
starts_with("Q6.1_"),
~ case_when(
. == 1 ~ 5,
. == 2 ~ 4,
. == 3 ~ 3,
. == 4 ~ 2,
. == 5 ~ 1,
TRUE ~ NA_real_
)),
manipulation_pass = case_when(
Q3.2 == 1 & Q7.1 == 3 ~ TRUE,
Q4.2 == 1 & Q7.1 == 1 ~ TRUE,
Q5.2 == 1 & Q7.1 == 2 ~ TRUE,
TRUE ~ FALSE
),
group = case_when(
Q3.2 == 1 ~ "Control",
Q4.2 == 1 ~ "Democracy",
Q5.2 == 1 ~ "Authoritarian",
TRUE ~ NA_character_
),
group_race = case_when(
Q11.2 == 1 ~ "Democracy_White",
Q12.2 == 1 ~ "Democracy_Asian",
Q13.2 == 1 ~ "Authoritarian_White",
Q14.2 == 1 ~ "Authoritarian_Asian",
TRUE ~ NA_character_
),
constitution = if_else(Q8.1 == 1, 1, 0),
ideology = Q8.2_1,
ideology_group = case_when(
ideology <= 3 ~ "Liberal",
ideology >= 4 & ideology <= 6 ~ "Neutral",
ideology >= 7 ~ "Conservative",
TRUE ~ NA_character_
),
psu = case_when(
Q2.1 %in% c(1, 4) ~ "In-partisans",
Q2.1 %in% 2:11 ~ "Out-partisans",
Q2.1 == 12 ~ "Independents",
TRUE ~ NA_character_
),
psu_rul = case_when(Q2.1 %in% c(1,4) ~ 1,
Q2.1 %in% 2:12 ~ 0,
TRUE ~ NA_real_),
psu_op = case_when(Q2.1 %in% 2:11 ~ 1,
Q2.1 %in% c(1,4,12)~ 0,
TRUE ~ NA_real_),
psu_indep = case_when(Q2.1 == 12 ~ 1,
Q2.1 %in% 1:11 ~ 0,
TRUE ~ NA_real_),
female = if_else(Q17.2 == 2, 1, 0),
age = Q17.3 + 17,
education = if_else(Q17.4 >= 5, NA_real_, as.numeric(Q17.4)),
income = Q17.5_1,
culture1 = case_when(Q8.5_1 == 1 ~ 5,
Q8.5_1 == 2 ~ 4,
Q8.5_1 == 3 ~ 3,
Q8.5_1 == 4 ~ 2,
Q8.5_1 == 5 ~ 1,
TRUE ~ NA_real_),
culture2 = case_when(Q8.5_2 == 1 ~ 5,
Q8.5_2 == 2 ~ 4,
Q8.5_2 == 3 ~ 3,
Q8.5_2 == 4 ~ 2,
Q8.5_2 == 5 ~ 1,
TRUE ~ NA_real_),
culture3 = case_when(Q8.5_3 == 1 ~ 5,
Q8.5_3 == 2 ~ 4,
Q8.5_3 == 3 ~ 3,
Q8.5_3 == 4 ~ 2,
Q8.5_3 == 5 ~ 1,
TRUE ~ NA_real_),
culture4 = case_when(Q8.5_4 == 1 ~ 5,
Q8.5_4 == 2 ~ 4,
Q8.5_4 == 3 ~ 3,
Q8.5_4 == 4 ~ 2,
Q8.5_4 == 5 ~ 1,
TRUE ~ NA_real_)
) %>%
rowwise() %>%
mutate(culture = mean(c(culture1, culture2, culture3, culture4), na.rm = FALSE)) %>%
ungroup() %>%
mutate(culture_group = if_else(culture < 12, "Neutral", "Superiority")) %>%
filter(!if_any(paste0("Q7_", 1:6), is.na)) %>%
{
fa_res <- fa(select(., Q7_1:Q7_6),
nfactors = 2,
rotate = "varimax",
fm = "ml")
scores <- factor.scores(select(., Q7_1:Q7_6), fa_res)$scores
bind_cols(., as_tibble(scores))
} %>%
mutate(
patr_group = cut(ML1,
breaks = c(-Inf, quantile(ML1, 0.33, na.rm = TRUE),
quantile(ML1, 0.67, na.rm = TRUE), Inf),
labels = c("Low", "Medium", "High"))
) %>%
dplyr::select(starts_with("Q6.1_"),
group, group_race,
constitution,
ideology, ideology_group,
psu, psu_rul, psu_op, psu_indep,
female, age, education, income,
culture, culture_group,
ML1, ML2, patr_group, manipulation_pass
)
security_counts <- read.csv("security.csv") %>%
mutate(
is_satisficer = (Q6.1_10 == 2),
group = case_when(
Q3.2 == 1 ~ "Control",
Q4.2 == 1 ~ "Democracy",
Q5.2 == 1 ~ "Authoritarian",
TRUE ~ NA_character_
)
)
summary_counts <- security_counts %>%
group_by(group) %>%
summarise(
total_n = n(),
n_satisficers = sum(is_satisficer, na.rm = TRUE),
n_dropped = total_n - n_satisficers
)
print(summary_counts)## # A tibble: 4 × 4
## group total_n n_satisficers n_dropped
## <chr> <int> <int> <int>
## 1 Authoritarian 749 550 199
## 2 Control 744 596 148
## 3 Democracy 702 518 184
## 4 <NA> 346 0 346Setting variables for the Study 2
taiwan <- read.csv("taiwan.csv") %>%
mutate(
across(
starts_with("Q7.1"),
~ case_when(
. == 1 ~ 5,
. == 2 ~ 4,
. == 4 ~ 2,
. == 5 ~ 1,
TRUE ~ .
)
),
across(
starts_with("Q7.2_"),
~ case_when(
. == 1 ~ 5,
. == 2 ~ 4,
. == 4 ~ 2,
. == 5 ~ 1,
TRUE ~ .
)
),
group = case_when(
Q3.2 == 1 ~ "Control",
Q4.2 == 1 ~ "Democracy",
Q5.2 == 1 ~ "Article9",
Q6.2 == 1 ~ "Democracy_Article9",
TRUE ~ NA_character_
),
knowledge = case_when(
Q8.1 == 1 ~ 1,
Q8.1 == 2 ~ 0,
Q8.1 %in% c(3, 4) ~ NA_real_,
TRUE ~ NA_real_
),
constitution = if_else(Q9.1 == 1, 1, 0),
ideology = Q9.2_1,
ideology_group = case_when(
Q9.2_1 <= 3 ~ "Liberal",
Q9.2_1 >= 4 & Q9.2_1 <= 6 ~ "Neutral",
Q9.2_1 >= 7 ~ "Conservative",
TRUE ~ NA_character_
),
psu = case_when(
Q2.1 %in% c(1,4) ~ "In-partisans",
Q2.1 %in% 2:11 ~ "Out-partisans",
Q2.1 == 12 ~ "Independents",
TRUE ~ NA_character_
),
psu_rul = case_when(Q2.1 %in% c(1,4) ~ 1, Q2.1 %in% 2:12 ~ 0, TRUE ~ NA_real_),
psu_op = case_when(Q2.1 %in% 2:11 ~ 1, Q2.1 %in% c(1,4,12) ~ 0, TRUE ~ NA_real_),
psu_indep = case_when(Q2.1 == 12 ~ 1, Q2.1 %in% 1:11 ~ 0, TRUE ~ NA_real_),
female = if_else(Q17.2 == 2, 1, 0),
age = Q17.3 + 17,
education = if_else(Q17.4 >= 5, NA_real_, as.numeric(Q17.4)),
income = Q17.5_1,
age_group = case_when(
age < 20 ~ "10s", age < 30 ~ "20s", age < 40 ~ "30s",
age < 50 ~ "40s", age < 60 ~ "50s", age < 70 ~ "60s", TRUE ~ "70s"
),
income_group = case_when(
income < 1000 ~ "0-1000",
income < 2000 ~ "1000-2000",
income <= 3000 ~ "2000-3000",
TRUE ~ ">3000"
)
) %>%
mutate(across(starts_with("Q9.5_"), ~ case_when(
. == 1 ~ 5, . == 2 ~ 4, . == 3 ~ 3, . == 4 ~ 2, . == 5 ~ 1, TRUE ~ NA_real_
), .names="culture_{col}")) %>%
rowwise() %>%
mutate(culture = mean(c_across(starts_with("culture_Q9.5_")), na.rm = FALSE)) %>%
ungroup() %>%
{
fa_res <- fa(select(., Q7_1:Q7_6), nfactors=1, fm="ml")
mutate(., patriotism = factor.scores(select(., Q7_1:Q7_6), fa_res)$scores[,1])
}
data <- taiwan %>%
filter(Q7.1_10 == 4) %>%
select(
group,
Q7.1_1, Q7.1_2, Q7.1_3, Q7.1_4, Q7.1_5, Q7.1_6,
Q7.1_7, Q7.1_8, Q7.1_9, Q7.1_10,
Q7.2_1, Q7.2_2, Q7.2_3, Q7.2_4, Q7.2_5, Q7.2_6,
Q7.2_7, Q7.2_8, Q7.2_9,
psu, psu_rul, psu_op, psu_indep,
ideology, ideology_group,
culture, constitution,
female, age, income, education,
knowledge, age_group, income_group
) %>%
drop_na()
summary_counts <- taiwan %>%
group_by(group) %>%
summarise(
total_n = n(),
n_satisficers = sum(Q7.1_10 != 4, na.rm = TRUE),
dif = total_n - n_satisficers
)
print(summary_counts)## # A tibble: 5 × 4
## group total_n n_satisficers dif
## <chr> <int> <int> <int>
## 1 Article9 534 121 413
## 2 Control 557 113 444
## 3 Democracy 554 126 428
## 4 Democracy_Article9 557 143 414
## 5 <NA> 342 0 342Code for the Main-Text Analyses
Table 1: Pass rate for manipulation check (Study 1)
check_summary <- security %>%
summarise(
N = n(),
Pass = sum(manipulation_pass, na.rm = TRUE),
Fail = N - Pass,
PassRate = round(Pass / N * 100, 1)
)N | Pass | Fail | PassRate |
|---|---|---|---|
1,651 | 787 | 864 | 47.7 |
Table 4: Descriptive statistics of Study 1
label_map2 <- c(
Q6.1_1 = "Military",
Q6.1_2 = "Comb.Weapon",
Q6.1_3 = "NoComb.Weapon",
Q6.1_4 = "Comb.Ammun.",
Q6.1_5 = "NoComb.Ammun.",
Q6.1_6 = "Comb.Trans.",
Q6.1_7 = "NoComb.Trans.",
Q6.1_8 = "Mil.Aid",
Q6.1_9 = "Fin.Aid",
psu_rul = "In-partisans",
psu_op = "Out-partisans",
psu_indep = "Independent",
ideology = "Ideology",
culture = "Cultural superiority",
ML1 = "Patriotism",
constitution= "Constitution",
female = "Female",
age = "Age",
income = "Income",
education = "Education"
)
desc_sec <- security %>%
select(all_of(names(label_map2))) %>%
summarise(across(everything(), list(
N = ~ sum(!is.na(.x)),
Mean = ~ round(mean(.x, na.rm = TRUE), 2),
SD = ~ round(sd(.x, na.rm = TRUE), 2),
Min = ~ round(min(.x, na.rm = TRUE), 2),
Max = ~ round(max(.x, na.rm = TRUE), 2),
Skew = ~ round(skewness(.x, na.rm = TRUE), 2)
), .names = "{col}__{fn}")) %>%
pivot_longer(
cols = everything(),
names_to = c("variable","stat"),
names_sep = "__",
values_to = "value"
) %>%
pivot_wider(
names_from = stat,
values_from = value
) %>%
mutate(
Variable = label_map2[variable]
) %>%
select(Variable, N, Mean, SD, Min, Max, Skew)Variable | N | Mean | SD | Min | Max | Skew |
|---|---|---|---|---|---|---|
Military | 1,651 | 1.90 | 1.08 | 1.00 | 5.00 | 0.89 |
Comb.Weapon | 1,651 | 2.19 | 1.11 | 1.00 | 5.00 | 0.53 |
NoComb.Weapon | 1,651 | 2.27 | 1.12 | 1.00 | 5.00 | 0.41 |
Comb.Ammun. | 1,651 | 2.17 | 1.09 | 1.00 | 5.00 | 0.54 |
NoComb.Ammun. | 1,651 | 2.27 | 1.11 | 1.00 | 5.00 | 0.37 |
Comb.Trans. | 1,651 | 2.17 | 1.12 | 1.00 | 5.00 | 0.55 |
NoComb.Trans. | 1,651 | 2.30 | 1.13 | 1.00 | 5.00 | 0.34 |
Mil.Aid | 1,651 | 2.40 | 1.13 | 1.00 | 5.00 | 0.26 |
Fin.Aid | 1,651 | 2.83 | 1.16 | 1.00 | 5.00 | -0.11 |
In-partisans | 1,441 | 0.22 | 0.41 | 0.00 | 1.00 | 1.37 |
Out-partisans | 1,441 | 0.38 | 0.49 | 0.00 | 1.00 | 0.48 |
Independent | 1,441 | 0.42 | 0.49 | 0.00 | 1.00 | 0.32 |
Ideology | 1,382 | 5.48 | 1.92 | 0.00 | 10.00 | 0.13 |
Cultural superiority | 1,648 | 3.33 | 0.77 | 1.00 | 5.00 | -0.16 |
Patriotism | 1,651 | 0.00 | 1.00 | -2.47 | 4.13 | 0.72 |
Constitution | 1,651 | 0.33 | 0.47 | 0.00 | 1.00 | 0.74 |
Female | 1,617 | 0.48 | 0.50 | 0.00 | 1.00 | 0.09 |
Age | 1,616 | 46.41 | 16.69 | 18.00 | 80.00 | 0.12 |
Income | 1,611 | 632.39 | 510.97 | 0.00 | 3,000.00 | 1.80 |
Education | 1,609 | 3.19 | 0.92 | 1.00 | 4.00 | -0.59 |
Table 5: Descriptive statics of Study 2
label_map <- c(
Q7.1_1 = "Military",
Q7.1_2 = "Comb.Weapon",
Q7.1_3 = "NoComb.Weapon",
Q7.1_4 = "Comb.Ammun.",
Q7.1_5 = "NoComb.Ammun.",
Q7.1_6 = "Comb.Trans.",
Q7.1_7 = "NoComb.Trans.",
Q7.1_8 = "Mil.Aid",
Q7.1_9 = "Fin.Aid",
psu_rul = "In-partisans",
psu_op = "Out-partisans",
psu_indep= "Independent",
ideology = "Ideology",
culture = "Cultural superiority",
constitution = "Constitution",
female = "Female",
age = "Age",
income = "Income",
education= "Education",
knowledge= "Knowledge on Taiwan"
)
desc_stats <- taiwan %>%
select(all_of(names(label_map))) %>%
summarise(across(everything(), list(
N = ~ sum(!is.na(.x)),
Mean = ~ round(mean(.x, na.rm = TRUE), 2),
SD = ~ round(sd(.x, na.rm = TRUE), 2),
Min = ~ round(min(.x, na.rm = TRUE), 2),
Max = ~ round(max(.x, na.rm = TRUE), 2),
Skew = ~ round(skewness(.x, na.rm = TRUE), 2)
), .names = "{col}__{fn}")) %>%
pivot_longer(
cols = everything(),
names_to = c("variable","stat"),
names_sep = "__",
values_to = "value"
) %>%
pivot_wider(
names_from = stat,
values_from = value
) %>%
mutate(
Variable = label_map[variable]
) %>%
select(Variable, N, Mean, SD, Min, Max, Skew)Variable | N | Mean | SD | Min | Max | Skew |
|---|---|---|---|---|---|---|
Military | 2,163 | 2.48 | 1.25 | 1 | 5 | 0.36 |
Comb.Weapon | 2,163 | 2.76 | 1.20 | 1 | 5 | 0.10 |
NoComb.Weapon | 2,163 | 2.86 | 1.19 | 1 | 5 | -0.03 |
Comb.Ammun. | 2,163 | 2.73 | 1.19 | 1 | 5 | 0.11 |
NoComb.Ammun. | 2,163 | 2.84 | 1.18 | 1 | 5 | -0.02 |
Comb.Trans. | 2,163 | 2.75 | 1.19 | 1 | 5 | 0.13 |
NoComb.Trans. | 2,163 | 2.87 | 1.19 | 1 | 5 | -0.01 |
Mil.Aid | 2,163 | 2.96 | 1.13 | 1 | 5 | -0.10 |
Fin.Aid | 2,163 | 3.25 | 1.12 | 1 | 5 | -0.33 |
In-partisans | 1,775 | 0.23 | 0.42 | 0 | 1 | 1.27 |
Out-partisans | 1,775 | 0.35 | 0.48 | 0 | 1 | 0.64 |
Independent | 1,775 | 0.45 | 0.50 | 0 | 1 | 0.21 |
Ideology | 1,616 | 5.42 | 1.95 | 0 | 10 | 0.09 |
Cultural superiority | 2,136 | 3.30 | 0.77 | 1 | 5 | -0.08 |
Constitution | 2,144 | 0.31 | 0.46 | 0 | 1 | 0.85 |
Female | 2,076 | 0.50 | 0.50 | 0 | 1 | 0.00 |
Age | 2,076 | 45.21 | 16.80 | 18 | 80 | 0.19 |
Income | 2,067 | 663.75 | 534.15 | 0 | 3,000 | 1.72 |
Education | 2,056 | 3.12 | 0.95 | 1 | 4 | -0.49 |
Knowledge on Taiwan | 1,075 | 0.83 | 0.37 | 0 | 1 | -1.79 |
Table 6: Balance check for Study 1
security <- security %>%
mutate(
age_group = case_when(
age < 20 ~ "10s",
age < 30 ~ "20s",
age < 40 ~ "30s",
age < 50 ~ "40s",
age < 60 ~ "50s",
age < 70 ~ "60s",
TRUE ~ ">70s"
),
income_group = case_when(
income < 1000 ~ "0–1000",
income < 2000 ~ "1000–2000",
income <= 3000 ~ "2000–3000",
TRUE ~ ">3000"
)
)
balance_vars <- c("psu_rul", "ideology_group", "age", "female", "education", "income_group")
tbl1 <- sumtable(
security,
group = "group",
vars = balance_vars,
out = "return",
group.test = TRUE
)
colnames(tbl1) <- make.unique(colnames(tbl1))Variable | N | Mean | SD | N.1 | Mean.1 | SD.1 | N.2 | Mean.2 | SD.2 | Test |
|---|---|---|---|---|---|---|---|---|---|---|
group | Authoritarian | Control | Democracy | |||||||
psu_rul | 474 | 0.22 | 0.42 | 516 | 0.23 | 0.42 | 451 | 0.19 | 0.4 | F=1.262 |
ideology_group | 455 | 495 | 432 | X2=2.766 | ||||||
... Conservative | 120 | 26% | 140 | 28% | 102 | 24% | ||||
... Liberal | 56 | 12% | 56 | 11% | 52 | 12% | ||||
... Neutral | 279 | 61% | 299 | 60% | 278 | 64% | ||||
age | 533 | 47 | 17 | 584 | 45 | 16 | 499 | 47 | 17 | F=2.509* |
female | 533 | 0.48 | 0.5 | 585 | 0.49 | 0.5 | 499 | 0.46 | 0.5 | F=0.407 |
education | 531 | 3.2 | 0.91 | 583 | 3.2 | 0.93 | 495 | 3.2 | 0.91 | F=0.089 |
income_group | 544 | 595 | 512 | X2=3.96 | ||||||
... >3000 | 11 | 2% | 13 | 2% | 16 | 3% | ||||
... 0–1000 | 433 | 80% | 482 | 81% | 411 | 80% | ||||
... 1000–2000 | 91 | 17% | 88 | 15% | 72 | 14% | ||||
... 2000–3000 | 9 | 2% | 12 | 2% | 13 | 3% |
Table 7: Balance check for Study 2
tbl2 <- sumtable(
data = taiwan,
group = "group",
vars = c("psu_rul","ideology_group","age","female","education","income_group","knowledge"),
out = "return",
group.test = TRUE
)
colnames(tbl2) <- make.unique(colnames(tbl2))Variable | N | Mean | SD | N.1 | Mean.1 | SD.1 | N.2 | Mean.2 | SD.2 | N.3 | Mean.3 | SD.3 | Test |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
group | Article9 | Control | Democracy | Democracy_Article9 | |||||||||
psu_rul | 432 | 0.22 | 0.42 | 452 | 0.21 | 0.41 | 444 | 0.26 | 0.44 | 432 | 0.24 | 0.43 | F=0.892 |
ideology_group | 381 | 429 | 415 | 391 | X2=8.352 | ||||||||
... Conservative | 104 | 27% | 99 | 23% | 101 | 24% | 108 | 28% | |||||
... Liberal | 35 | 9% | 60 | 14% | 49 | 12% | 54 | 14% | |||||
... Neutral | 242 | 64% | 270 | 63% | 265 | 64% | 229 | 59% | |||||
age | 501 | 45 | 17 | 533 | 46 | 16 | 516 | 45 | 17 | 526 | 45 | 17 | F=0.65 |
female | 501 | 0.53 | 0.5 | 533 | 0.49 | 0.5 | 516 | 0.48 | 0.5 | 526 | 0.5 | 0.5 | F=0.764 |
education | 496 | 3.1 | 0.96 | 529 | 3.2 | 0.93 | 513 | 3.1 | 0.94 | 518 | 3.1 | 0.97 | F=0.801 |
income_group | 534 | 557 | 554 | 557 | X2=7.696 | ||||||||
... >3000 | 35 | 7% | 27 | 5% | 41 | 7% | 32 | 6% | |||||
... 0-1000 | 404 | 76% | 436 | 78% | 407 | 73% | 413 | 74% | |||||
... 1000-2000 | 77 | 14% | 82 | 15% | 88 | 16% | 96 | 17% | |||||
... 2000-3000 | 18 | 3% | 12 | 2% | 18 | 3% | 16 | 3% | |||||
knowledge | 228 | 0.78 | 0.41 | 236 | 0.83 | 0.37 | 351 | 0.87 | 0.33 | 260 | 0.82 | 0.38 | F=3.037** |
Figure 3: Distributions of each outcome variable within the control group (Study 1)
data2 <- security %>%
filter(group == "Control") %>%
pivot_longer(
cols = starts_with("Q6.1_"),
names_to = "Item",
values_to = "Value"
) %>%
mutate(
Item_JP = case_when(
Item == "Q6.1_1" ~ "Military",
Item == "Q6.1_2" ~ "Comb.\nWeapon",
Item == "Q6.1_3" ~ "NoComb.\nWeapon",
Item == "Q6.1_4" ~ "Comb.\nAmmun.",
Item == "Q6.1_5" ~ "NoComb.\nAmmun.",
Item == "Q6.1_6" ~ "Comb.\nTrans.",
Item == "Q6.1_7" ~ "NoComb.\nTrans.",
Item == "Q6.1_8" ~ "Mil.\nAid",
Item == "Q6.1_9" ~ "Fin.\nAid",
TRUE ~ NA_character_
),
Value = factor(
Value,
levels = 1:5,
labels = c(
"Strongly disagree",
"Somewhat disagree",
"Neither",
"Somewhat agree",
"Agree"
)
)
) %>%
filter(!is.na(Item_JP)) %>%
mutate(Item_JP = factor(
Item_JP,
levels = c(
"Military",
"Comb.\nWeapon",
"NoComb.\nWeapon",
"Comb.\nAmmun.",
"NoComb.\nAmmun.",
"Comb.\nTrans.",
"NoComb.\nTrans.",
"Mil.\nAid",
"Fin.\nAid"
)
))
hist_df <- data2 %>%
filter(!is.na(Value)) %>%
group_by(Item_JP, Value) %>%
summarise(count = n(), .groups="drop") %>%
group_by(Item_JP) %>%
mutate(pct = count / sum(count) * 100) %>%
ungroup()
ggplot(hist_df, aes(x = Value, y = pct)) +
geom_col(fill = "grey60") +
geom_label(
aes(label = paste0(round(pct, 1), "%")),
fill = "white",
label.size = 0.2,
size = 3,
vjust = -0.2
) +
facet_wrap(~ Item_JP, scales = "free_y") +
scale_y_continuous(
expand = expansion(mult = c(0, 0.1)),
labels = scales::percent_format(scale = 1)
) +
labs(
x = "Agree / Disagree",
y = "(%)",
title = ""
) +
theme_bw(base_size = 12) +
theme(
axis.text.x = element_text(angle = 45, vjust = 0.5),
strip.text = element_text(size = 10),
panel.grid.major.x = element_blank()
)
Figure 4: Distributions of each outcome variable within the control group (Study 2)
data2 <- taiwan %>%
filter(group == "Control") %>%
pivot_longer(
cols = starts_with("Q7.1_"),
names_to = "Item",
values_to = "Value"
) %>%
mutate(
Item_JP = case_when(
Item == "Q7.1_1" ~ "Military",
Item == "Q7.1_2" ~ "Comb.\nWeapon",
Item == "Q7.1_3" ~ "NoComb.\nWeapon",
Item == "Q7.1_4" ~ "Comb.\nAmmun.",
Item == "Q7.1_5" ~ "NoComb.\nAmmun.",
Item == "Q7.1_6" ~ "Comb.\nTrans.",
Item == "Q7.1_7" ~ "NoComb.\nTrans.",
Item == "Q7.1_8" ~ "Mil.\nAid",
Item == "Q7.1_9" ~ "Fin.\nAid",
TRUE ~ NA_character_
),
Value = factor(
Value,
levels = 1:5,
labels = c(
"Strongly disagree",
"Somewhat disagree",
"Neither",
"Somewhat agree",
"Agree"
)
)
) %>%
filter(!is.na(Item_JP)) %>%
mutate(Item_JP = factor(
Item_JP,
levels = c(
"Military",
"Comb.\nWeapon",
"NoComb.\nWeapon",
"Comb.\nAmmun.",
"NoComb.\nAmmun.",
"Comb.\nTrans.",
"NoComb.\nTrans.",
"Mil.\nAid",
"Fin.\nAid"
)
))
hist_df <- data2 %>%
filter(!is.na(Value)) %>%
group_by(Item_JP, Value) %>%
summarise(count = n(), .groups="drop") %>%
group_by(Item_JP) %>%
mutate(pct = count / sum(count) * 100) %>%
ungroup()
ggplot(hist_df, aes(x = Value, y = pct)) +
geom_col(fill = "grey60") +
geom_label(
aes(label = paste0(round(pct, 1), "%")),
fill = "white",
label.size = 0.2,
size = 3,
vjust = -0.2
) +
facet_wrap(~ Item_JP, scales = "free_y") +
scale_y_continuous(
expand = expansion(mult = c(0, 0.1)),
labels = scales::percent_format(scale = 1)
) +
labs(
x = "Agree / Disagree",
y = "(%)",
title = ""
) +
theme_bw(base_size = 12) +
theme(
axis.text.x = element_text(angle = 45, vjust = 0.5),
strip.text = element_text(size = 10),
panel.grid.major.x = element_blank()
)
Figure 5: Mean difference of each outcome variable (Study 1)
data2 <- security %>%
select(
group,
starts_with("Q6.1_"),
psu, psu_rul, psu_op, psu_indep,
ideology, ideology_group,
patr_group, ML1, culture, culture_group,
constitution,
female, age, income, education
) %>%
drop_na()
stats0 <- data2 %>%
pivot_longer(
cols = starts_with("Q6.1_"),
names_to = "Item",
values_to = "Value"
) %>%
group_by(Item, group) %>%
summarise(
n = n(),
mean = mean(Value, na.rm = TRUE),
sd = sd(Value, na.rm = TRUE),
se = sd / sqrt(n),
.groups = "drop"
)
ctrl_stats <- stats0 %>%
filter(group == "Control") %>%
select(Item, ctrl_mean = mean, ctrl_se = se)
stats <- stats0 %>%
left_join(ctrl_stats, by = "Item") %>%
mutate(
diff = mean - ctrl_mean,
se_diff = sqrt(se^2 + ctrl_se^2),
lower_diff = diff - 1.96 * se_diff,
upper_diff = diff + 1.96 * se_diff,
signif = (lower_diff > 0) | (upper_diff < 0),
label = paste0(round(diff, 2), "\n[",
round(lower_diff, 2), ", ",
round(upper_diff, 2), "]"),
Item_JP = case_when(
Item == "Q6.1_1" ~ "Military",
Item == "Q6.1_2" ~ "Comb.\nWeapon",
Item == "Q6.1_3" ~ "NoComb.\nWeapon",
Item == "Q6.1_4" ~ "Comb.\nAmmun.",
Item == "Q6.1_5" ~ "NoComb.\nAmmun.",
Item == "Q6.1_6" ~ "Comb.\nTrans.",
Item == "Q6.1_7" ~ "NoComb.\nTrans.",
Item == "Q6.1_8" ~ "Mil.\nAid",
Item == "Q6.1_9" ~ "Fin.\nAid",
TRUE ~ NA_character_
)
) %>%
filter(group %in% c("Democracy", "Authoritarian") & !is.na(Item_JP)) %>%
mutate(
group = factor(group, levels = c("Democracy", "Authoritarian")),
Item_JP = factor(Item_JP,
levels = c(
"Military", "Comb.\nWeapon", "NoComb.\nWeapon",
"Comb.\nAmmun.", "NoComb.\nAmmun.",
"Comb.\nTrans.", "NoComb.\nTrans.",
"Mil.\nAid", "Fin.\nAid"
))
)
pd <- position_dodge(width = 0.6)
ggplot(stats, aes(x = group, y = diff, ymin = lower_diff, ymax = upper_diff)) +
geom_errorbar(data = filter(stats, !signif),
aes(group = group),
position = pd, color = "grey70", size = 0.6, width = 0.2) +
geom_point(data = filter(stats, !signif),
aes(group = group),
position = pd, color = "grey70", size = 2) +
geom_errorbar(data = filter(stats, signif),
aes(group = group),
position = pd, color = "black", size = 1.2, width = 0.2) +
geom_point(data = filter(stats, signif),
aes(group = group),
position = pd, color = "black", size = 3) +
geom_text(data = filter(stats, signif),
aes(label = label, group = group),
position = pd,
hjust = -0.2,
size = 3) +
facet_wrap(~ Item_JP, scales = "free_y") +
theme_bw(base_size = 12) +
theme(
axis.text.x = element_text(angle = 45, hjust = 1),
strip.text = element_text(size = 10)
) +
labs(
x = "Group (vs. Control)",
y = "Difference from Control",
title = ""
)
Figure 6: Ordinary Least Squares estimation’ coefficients plot (Study 1)
data_ols <- security %>%
mutate(
Democracy = as.integer(group == "Democracy"),
Authoritarian = as.integer(group == "Authoritarian")
) %>%
select(
starts_with("Q6.1_"),
Democracy, Authoritarian,
psu_rul, ideology, ML1, culture,
constitution, female, age, income, education
) %>%
drop_na()
models1 <- lapply(paste0("Q6.1_", 1:5), function(outcome) {
lm(
reformulate(
c("Democracy","Authoritarian","psu_rul","ideology",
"ML1","culture","constitution","female","age","income","education"),
response = outcome
),
data = data_ols
)
})
models2 <- lapply(paste0("Q6.1_", 6:9), function(outcome) {
lm(
reformulate(
c("Democracy","Authoritarian","psu_rul","ideology",
"ML1","culture","constitution","female","age","income","education"),
response = outcome
),
data = data_ols
)
})
cov_labels <- c(
"Democracy","Authoritarian","Partisanship","Ideology",
"Patriotism","Culture","Constitution",
"Female","Age","Income","Education"
)
all_models <- c(models1, models2)
outcomes <- c("Military","Comb.Weapon","NoComb.Weapon","Comb.Ammun.","NoComb.Ammun.",
"Comb.Trans.","NoComb.Trans.","Mil.Aid","Fin.Aid")
names(all_models) <- outcomes
coef_df <- imap_dfr(all_models, ~ {
tidy(.x) %>%
filter(term %in% c("Democracy","Authoritarian")) %>%
transmute(
outcome = .y,
term,
estimate,
se = std.error,
p.value,
lower = estimate - 1.96*std.error,
upper = estimate + 1.96*std.error,
sig = p.value < 0.05,
label = ifelse(sig,
paste0(round(estimate,2), "\n[",
round(lower,2), ",", round(upper,2), "]"),
NA_character_)
)
}) %>%
mutate(
term_jp = recode(term,
Democracy = "Democracy",
Authoritarian = "Authoritarian"
),
term_jp = factor(term_jp, levels = c("Democracy","Authoritarian")),
outcome = factor(outcome, levels = outcomes)
)
ggplot(coef_df, aes(x = estimate, y = term_jp)) +
geom_vline(xintercept = 0, linetype = "dashed") +
geom_errorbarh(data = filter(coef_df, !sig),
aes(xmin = lower, xmax = upper),
color = "grey80", height = 0) +
geom_point(data = filter(coef_df, !sig),
color = "grey80", size = 2) +
geom_errorbarh(data = filter(coef_df, sig),
aes(xmin = lower, xmax = upper),
color = "black", size = 1, height = 0) +
geom_point(data = filter(coef_df, sig),
color = "black", size = 3) +
geom_text(data = filter(coef_df, sig),
aes(label = label),
nudge_y = 0.2, size = 3) +
facet_wrap(~ outcome, ncol = 3) +
theme_bw(base_size = 12) +
theme(
strip.text = element_text(size = 10),
axis.title.y = element_blank()
) +
labs(
x = "Estimated Treatment Effect",
title = ""
)
Figure 7: Predicted values of military intervention by moderators (Study 1: Constitutional understanding; Cultural superiority; Patriotism; In-partisanship)
Figure 8: Predicted values of financial aids by moderators (Study 1: Constitutional understanding; Cultural superiority; Patriotism; In-partisanship)
data_ols <- security %>%
mutate(
treatment = factor(group,
levels = c("Control","Democracy","Authoritarian"))
)
mod_names <- c(
constitution = "Constitution",
culture = "Culture",
ML1 = "Patriotism",
psu_rul = "In-partisans"
)
make_plot <- function(dv, modvar) {
fmla <- as.formula(paste0(
dv, " ~ treatment * ", modvar,
" + ideology + female + age + income + education"
))
fit <- lm(fmla, data = data_ols)
visreg(
fit,
xvar = modvar,
by = "treatment",
overlay = TRUE,
gg = TRUE,
line = list(size = 1.0),
fill = list(alpha = 0.2)
) +
labs(
x = mod_names[modvar],
y = "Predicted Support for Military"
) +
scale_color_discrete(
name = "Treatment",
labels = c("Control","Democracy","Article9","Democracy+9")
) +
scale_fill_discrete(guide = "none") +
theme_bw(base_size = 12) +
theme(
legend.position = "bottom",
legend.direction = "horizontal",
legend.title = element_text(size = 10),
legend.text = element_text(size = 8),
legend.margin = margin(t = -2, r = 0, b = 0, l = 0),
legend.box.margin = margin(0, 0, 0, 0)
)
}
mods <- c("constitution","culture","ML1","psu_rul")
plots_mil <- lapply(mods, function(m) make_plot("Q6.1_1", m))
legend_mil <- cowplot::get_legend(
plots_mil[[1]] +
theme(
legend.position = "bottom",
legend.direction = "horizontal"
)
)
plots_mil_noleg <- lapply(plots_mil, function(p) p + theme(legend.position = "none"))
grid_mil <- wrap_plots(plots_mil_noleg, ncol = 2)
fig_mil <- cowplot::plot_grid(
grid_mil,
legend_mil,
ncol = 1,
rel_heights = c(1, 0.10)
)
print(fig_mil)
plots_finaid <- lapply(mods, function(m) make_plot("Q6.1_9", m))
legend_finaid <- cowplot::get_legend(plots_finaid[[1]] + theme(legend.position = "bottom"))
plots_finaid_noleg <- lapply(plots_finaid, function(p) p + theme(legend.position = "none"))
grid_finaid <- wrap_plots(plots_finaid_noleg, ncol = 2)
fig_finaid <- cowplot::plot_grid(grid_finaid, legend_finaid, ncol = 1, rel_heights = c(1, 0.10))
print(fig_finaid)
Figure 9: Mean difference of each outcome variable (Study 2)
labels <- c(
"Q7.1_1" = "Military", "Q7.1_2" = "Comb.Weapon",
"Q7.1_3" = "NoComb.Weapon", "Q7.1_4" = "Comb.Ammun.",
"Q7.1_5" = "NoComb.Ammun.", "Q7.1_6" = "Comb.Trans.",
"Q7.1_7" = "NoComb.Trans.", "Q7.1_8" = "Mil.Aid",
"Q7.1_9" = "Fin.Aid"
)
label_order <- c(
"Military", "Comb.Weapon", "NoComb.Weapon",
"Comb.Ammun.", "NoComb.Ammun.",
"Comb.Trans.", "NoComb.Trans.",
"Mil.Aid", "Fin.Aid"
)
group_order <- c("Democracy", "Article9", "Democracy_Article9", "Control")
stats0 <- taiwan %>%
pivot_longer(
cols = paste0("Q7.1_", 1:9),
names_to = "Item",
values_to = "Value"
) %>%
group_by(Item, group) %>%
summarise(
n = n(),
mean = mean(Value, na.rm = TRUE),
sd = sd(Value, na.rm = TRUE),
se = sd / sqrt(n),
.groups = "drop"
)
ctrl_stats <- stats0 %>%
filter(group == "Control") %>%
select(Item, ctrl_mean = mean, ctrl_se = se)
stats <- stats0 %>%
left_join(ctrl_stats, by = "Item") %>%
mutate(
diff = mean - ctrl_mean,
se_diff = sqrt(se^2 + ctrl_se^2),
lower_diff = diff - 1.96 * se_diff,
upper_diff = diff + 1.96 * se_diff,
signif = (lower_diff > 0) | (upper_diff < 0),
label = paste0(
round(diff, 2),
" [", round(lower_diff, 2), ",", round(upper_diff, 2), "]"
),
Item_JP = case_when(
Item == "Q7.1_1" ~ "Military",
Item == "Q7.1_2" ~ "Comb.\nWeapon",
Item == "Q7.1_3" ~ "NoComb.\nWeapon",
Item == "Q7.1_4" ~ "Comb.\nAmmun.",
Item == "Q7.1_5" ~ "NoComb.\nAmmun.",
Item == "Q7.1_6" ~ "Comb.\nTrans.",
Item == "Q7.1_7" ~ "NoComb.\nTrans.",
Item == "Q7.1_8" ~ "Mil.\nAid",
Item == "Q7.1_9" ~ "Fin.\nAid",
TRUE ~ NA_character_
)
) %>%
filter(group != "Control") %>%
mutate(
group = factor(group, levels = c("Democracy", "Article9", "Democracy_Article9")),
Item_JP = factor(
Item_JP,
levels = c(
"Military",
"Comb.\nWeapon",
"NoComb.\nWeapon",
"Comb.\nAmmun.",
"NoComb.\nAmmun.",
"Comb.\nTrans.",
"NoComb.\nTrans.",
"Mil.\nAid",
"Fin.\nAid"
)
)
)
ggplot() +
geom_errorbar(
data = filter(stats, !signif),
aes(x = group, ymin = lower_diff, ymax = upper_diff),
color = "grey70", size = 0.5, width = 0.2
) +
geom_point(
data = filter(stats, !signif),
aes(x = group, y = diff),
color = "grey70", size = 2
) +
geom_errorbar(
data = filter(stats, signif),
aes(x = group, ymin = lower_diff, ymax = upper_diff),
color = "black", size = 1.2, width = 0.2
) +
geom_point(
data = filter(stats, signif),
aes(x = group, y = diff),
color = "black", size = 3
) +
geom_text(
data = filter(stats, signif),
aes(x = group, y = upper_diff, label = label),
vjust = -0.5, size = 3
) +
facet_wrap(~ Item_JP, scales = "free_y") +
theme_bw(base_size = 12) +
theme(
axis.text.x = element_text(angle = 45, hjust = 1),
strip.text = element_text(size = 10)
) +
labs(
x = "Group (vs. Control)",
y = "Difference from Control"
)Figure 10: Ordinary Least Squares estimation’ coefficients plot (Study 2)
data_ols <- taiwan %>%
mutate(
Article9 = as.integer(group == "Article9"),
Democracy = as.integer(group == "Democracy"),
Democracy_Article9 = as.integer(group == "Democracy_Article9")
)
run_ols <- function(items, file, col_labels) {
models <- lapply(items, function(outcome) {
lm(
formula = as.formula(
paste0(outcome,
" ~ Article9 + Democracy + Democracy_Article9 + psu_rul + ideology + culture + constitution + knowledge + female + age + income + education")
),
data = data_ols
)
})
stargazer(
models,
column.labels = col_labels,
covariate.labels = c(
"Article9",
"Democracy",
"DemocracyArticle9",
"Partisanship",
"Ideology",
"Culture",
"Constitution",
"Knowledge",
"Female",
"Age",
"Income",
"Education"
),
type = "text",
out = file
)
}
models1 <- lapply(paste0("Q7.1_", 1:5), function(outcome) {
lm(as.formula(
paste0(outcome,
" ~ Article9 + Democracy + Democracy_Article9 + psu_rul + ideology + culture + constitution + knowledge + female + age + income + education")
),
data = data_ols)
})
models2 <- lapply(paste0("Q7.1_", 6:9), function(outcome) {
lm(as.formula(
paste0(outcome,
" ~ Article9 + Democracy + Democracy_Article9 + psu_rul + ideology + culture + constitution + knowledge + female + age + income + education")
),
data = data_ols)
})
models1_q72 <- lapply(paste0("Q7.2_", 1:5), function(outcome) {
lm(as.formula(
paste0(outcome,
" ~ Article9 + Democracy + Democracy_Article9 + ",
"psu_rul + ideology + culture + constitution + knowledge + ",
"female + age + income + education")
),
data = data_ols)
})
models2_q72 <- lapply(paste0("Q7.2_", 6:9), function(outcome) {
lm(as.formula(
paste0(outcome,
" ~ Article9 + Democracy + Democracy_Article9 + ",
"psu_rul + ideology + culture + constitution + knowledge + ",
"female + age + income + education")
),
data = data_ols)
})
all_models <- c(models1, models2)
outcomes <- c("Military","Comb.Weapon","NoComb.Weapon","Comb.Ammun.","NoComb.Ammun.",
"Comb.Trans.","NoComb.Trans.","Mil.Aid","Fin.Aid")
names(all_models) <- outcomes
coef_df <- imap_dfr(all_models, ~ {
broom::tidy(.x) %>%
filter(term %in% c("Article9","Democracy","Democracy_Article9")) %>%
transmute(
outcome = .y,
term,
estimate,
std.error,
p.value,
lower = estimate - 1.96*std.error,
upper = estimate + 1.96*std.error,
sig = p.value < 0.05
)
}) %>%
mutate(
term_jp = case_when(
term == "Article9" ~ "Article9",
term == "Democracy" ~ "Democracy",
term == "Democracy_Article9" ~ "Democracy+Article9"
),
term_jp = factor(term_jp, levels = c("Article9","Democracy","Democracy+Article9")),
outcome = factor(outcome, levels = outcomes),
label = ifelse(
sig,
paste0(
round(estimate,2),
"\n[", round(lower,2), ",", round(upper,2), "]"
),
NA_character_
)
)
pd <- position_dodge(width = 0.6)
ggplot(coef_df, aes(x = estimate, y = term_jp)) +
geom_vline(xintercept = 0, linetype = "dashed", color = "black") +
geom_errorbarh(
data = filter(coef_df, !sig),
aes(xmin = lower, xmax = upper),
height = 0,
color = "grey80"
) +
geom_point(
data = filter(coef_df, !sig),
color = "grey80",
size = 2
) +
geom_errorbarh(
data = filter(coef_df, sig),
aes(xmin = lower, xmax = upper),
height = 0,
color = "black",
size = 1
) +
geom_point(
data = filter(coef_df, sig),
color = "black",
size = 3
) +
geom_text(
data = filter(coef_df, sig),
aes(label = label),
nudge_y = 0.2,
size = 3
) +
facet_wrap(~ outcome, ncol = 3) +
theme_bw(base_size = 12) +
theme(
strip.text = element_text(size = 10),
axis.title.y = element_blank()
) +
labs(
x = "Estimated Treatment Effect",
title = ""
)
Figure 11: Predicted values of military intervention by moderators (Study 2: Constitutional understanding; Cultural superiority; Knowledge on Taiwan political system; In-partisanship)
Figure 12: Predicted values of financial aids by moderators (Study 2: Constitutional understanding; Cultural superiority; Knowledge on Taiwan political system; In-partisanship)
mod_names <- c(
constitution = "Constitution",
culture = "Culture",
knowledge = "Knowledge",
psu_rul = "In-partisans"
)
data_ols <- data_ols %>%
mutate(
treatment = factor(
group,
levels = c("Control","Democracy","Article9","Democracy_Article9")
)
)
make_plot <- function(dv, modvar) {
fmla <- as.formula(paste0(
dv, " ~ treatment * ", modvar,
" + ideology + female + age + income + education"
))
fit <- lm(fmla, data = data_ols)
visreg(
fit,
xvar = modvar,
by = "treatment",
overlay = TRUE,
gg = TRUE,
line = list(size = 1.2),
fill = list(alpha = 0.2)
) +
labs(
x = mod_names[modvar],
y = if (dv == "Q7.1_1") "Predicted Support for Military"
else "Predicted Support for Financial Aid"
) +
scale_color_discrete(
name = "Treatment",
labels = c("Control","Democracy","Article9","Democracy+9")
) +
scale_fill_discrete(guide = "none") +
theme_bw(base_size = 12) +
theme(
legend.position = "bottom",
legend.title = element_text(size = 10),
legend.text = element_text(size = 8),
legend.direction = "horizontal",
legend.margin = margin(t = -2, r = 0, b = 0, l = 0),
legend.box.margin = margin(0, 0, 0, 0)
)
}
mods <- c("constitution","culture","knowledge","psu_rul")
plots_mil <- lapply(mods, function(m) make_plot("Q7.1_1", m))
plots_finaid <- lapply(mods, function(m) make_plot("Q7.1_9", m))
legend_mil <- cowplot::get_legend(
plots_mil[[1]] +
theme(legend.position = "bottom",
legend.direction = "horizontal")
)
plots_mil_noleg <- lapply(plots_mil, function(p) p + theme(legend.position = "none"))
grid_mil <- wrap_plots(plots_mil_noleg, ncol = 2)
figure_mil <- cowplot::plot_grid(
grid_mil,
legend_mil,
ncol = 1,
rel_heights = c(1, 0.10)
)
print(figure_mil) 
legend_fa <- cowplot::get_legend(
plots_finaid[[1]] +
theme(legend.position = "bottom",
legend.direction = "horizontal")
)
plots_finaid_noleg <- lapply(plots_finaid, function(p) p + theme(legend.position = "none"))
grid_fa <- wrap_plots(plots_finaid_noleg, ncol = 2)
figure_fa <- cowplot::plot_grid(
grid_fa,
legend_fa,
ncol = 1,
rel_heights = c(1, 0.10)
)
print(figure_fa) 
Code for the Online Appendix
A. All of Ordinary Least Squares Results
Table A1: Ordinary Least Squares results for Study 1 (For five dependent variables)
Table A2: Ordinary Least Squares results for Study 1 (For four dependent variables)
data_ols <- security %>%
mutate(
Democracy = as.integer(group == "Democracy"),
Authoritarian = as.integer(group == "Authoritarian")
) %>%
select(
starts_with("Q6.1_"),
Democracy, Authoritarian,
psu_rul, ideology, ML1, culture,
constitution, female, age, income, education
) %>%
drop_na()
models1 <- lapply(paste0("Q6.1_", 1:5), function(outcome) {
lm(
reformulate(
c("Democracy","Authoritarian","psu_rul","ideology",
"ML1","culture","constitution","female","age","income","education"),
response = outcome
),
data = data_ols
)
})
models2 <- lapply(paste0("Q6.1_", 6:9), function(outcome) {
lm(
reformulate(
c("Democracy","Authoritarian","psu_rul","ideology",
"ML1","culture","constitution","female","age","income","education"),
response = outcome
),
data = data_ols
)
})
cov_labels <- c(
"Democracy","Authoritarian","Partisanship","Ideology",
"Patriotism","Culture","Constitution",
"Female","Age","Income","Education"
)| Military | Comb.Weapon | NoComb.Weapon | Comb.Ammun. | NoComb.Ammun. | |
|---|---|---|---|---|---|
| Democracy | 0.144* (0.074) | 0.256*** (0.077) | 0.322*** (0.078) | 0.227*** (0.076) | 0.253*** (0.078) |
| Authoritarian | -0.028 (0.073) | -0.055 (0.075) | -0.048 (0.076) | -0.085 (0.075) | -0.060 (0.076) |
| Partisanship | 0.199*** (0.073) | 0.168** (0.076) | 0.212*** (0.077) | 0.244*** (0.075) | 0.233*** (0.077) |
| Ideology | 0.005 (0.016) | 0.012 (0.016) | 0.002 (0.016) | 0.007 (0.016) | -0.011 (0.016) |
| Patriotism | 0.101*** (0.034) | 0.075** (0.035) | 0.072** (0.036) | 0.047 (0.035) | 0.041 (0.036) |
| Culture | 0.176*** (0.042) | 0.236*** (0.044) | 0.206*** (0.044) | 0.182*** (0.044) | 0.172*** (0.044) |
| Constitution | 0.242*** (0.067) | 0.284*** (0.069) | 0.353*** (0.070) | 0.326*** (0.069) | 0.370*** (0.070) |
| Female | -0.158** (0.063) | -0.290*** (0.065) | -0.253*** (0.066) | -0.235*** (0.065) | -0.256*** (0.066) |
| Age | -0.010*** (0.002) | -0.010*** (0.002) | -0.010*** (0.002) | -0.010*** (0.002) | -0.009*** (0.002) |
| Income | 0.000** (0.000) | 0.000 (0.000) | 0.000 (0.000) | 0.000* (0.000) | 0.000 (0.000) |
| Education | -0.077** (0.034) | -0.077** (0.035) | -0.070* (0.036) | -0.059* (0.035) | -0.063* (0.036) |
| Observations | 1207 | 1207 | 1207 | 1207 | 1207 |
| R-squared | 0.083 | 0.110 | 0.114 | 0.109 | 0.103 |
| F-statistic | 9.804 | 13.493 | 13.983 | 13.278 | 12.532 |
| ✱ Signif. codes: * p<0.10, ** p<0.05, *** p<0.01. Standard errors in parentheses. |
| Comb.Trans. | NoComb.Trans. | Mil.Aid | Fin.Aid | |
|---|---|---|---|---|
| Democracy | 0.236*** (0.078) | 0.290*** (0.080) | 0.304*** (0.079) | 0.265*** (0.083) |
| Authoritarian | -0.053 (0.076) | -0.040 (0.078) | -0.014 (0.078) | -0.152* (0.081) |
| Partisanship | 0.238*** (0.077) | 0.208*** (0.079) | 0.175** (0.078) | 0.340*** (0.082) |
| Ideology | 0.009 (0.016) | -0.010 (0.017) | -0.029* (0.017) | -0.051*** (0.017) |
| Patriotism | 0.039 (0.036) | 0.048 (0.037) | 0.057 (0.037) | -0.074* (0.038) |
| Culture | 0.175*** (0.044) | 0.172*** (0.046) | 0.212*** (0.045) | 0.140*** (0.048) |
| Constitution | 0.326*** (0.070) | 0.347*** (0.072) | 0.305*** (0.071) | 0.197*** (0.075) |
| Female | -0.307*** (0.066) | -0.267*** (0.068) | -0.189*** (0.068) | -0.136* (0.071) |
| Age | -0.010*** (0.002) | -0.007*** (0.002) | -0.010*** (0.002) | -0.003 (0.002) |
| Income | 0.000 (0.000) | 0.000 (0.000) | 0.000 (0.000) | -0.000 (0.000) |
| Education | -0.047 (0.036) | -0.060 (0.037) | -0.096*** (0.037) | -0.030 (0.038) |
| Observations | 1207 | 1207 | 1207 | 1207 |
| R-squared | 0.109 | 0.095 | 0.092 | 0.072 |
| F-statistic | 13.290 | 11.395 | 10.975 | 8.395 |
| ✱ Signif. codes: * p<0.10, ** p<0.05, *** p<0.01. Standard errors in parentheses. |
Table A3: Ordinary Least Squares results for Study 2 (For five dependent variables)
Table A4: Ordinary Least Squares results for Study 2 (For four dependent variables)
data_ols <- taiwan %>%
mutate(
Article9 = as.integer(group == "Article9"),
Democracy = as.integer(group == "Democracy"),
Democracy_Article9 = as.integer(group == "Democracy_Article9")
)
models1 <- lapply(paste0("Q7.1_", 1:5), function(outcome) {
lm(as.formula(
paste0(outcome,
" ~ Article9 + Democracy + Democracy_Article9 + psu_rul + ideology + culture + constitution + knowledge + female + age + income + education")
),
data = data_ols)
})
models2 <- lapply(paste0("Q7.1_", 6:9), function(outcome) {
lm(as.formula(
paste0(outcome,
" ~ Article9 + Democracy + Democracy_Article9 + psu_rul + ideology + culture + constitution + knowledge + female + age + income + education")
),
data = data_ols)
})| Military | Comb.Weapon | NoComb.Weapon | Comb.Ammun. | NoComb.Ammun. | |
|---|---|---|---|---|---|
| Article9 | 0.216 (0.132) | 0.205 (0.127) | 0.096 (0.123) | 0.056 (0.127) | 0.017 (0.123) |
| Democracy | 0.007 (0.121) | -0.024 (0.117) | -0.151 (0.113) | -0.131 (0.117) | -0.146 (0.114) |
| DemocracyArticle9 | 0.243* (0.129) | 0.247** (0.124) | 0.211* (0.120) | 0.184 (0.124) | 0.228* (0.120) |
| Partisanship | 0.157 (0.100) | 0.227** (0.096) | 0.258*** (0.094) | 0.186* (0.097) | 0.285*** (0.094) |
| Ideology | 0.011 (0.022) | -0.005 (0.021) | 0.006 (0.021) | -0.005 (0.021) | 0.014 (0.021) |
| Culture | 0.312*** (0.058) | 0.267*** (0.056) | 0.217*** (0.054) | 0.240*** (0.056) | 0.191*** (0.054) |
| Constitution | 0.773*** (0.093) | 0.802*** (0.089) | 0.721*** (0.087) | 0.757*** (0.089) | 0.697*** (0.087) |
| Knowledge | -0.147 (0.125) | 0.024 (0.120) | 0.058 (0.117) | 0.045 (0.120) | 0.141 (0.117) |
| Female | -0.415*** (0.095) | -0.460*** (0.091) | -0.469*** (0.088) | -0.440*** (0.091) | -0.509*** (0.089) |
| Age | -0.004* (0.003) | -0.004 (0.003) | -0.003 (0.002) | -0.003 (0.003) | -0.002 (0.002) |
| Income | 0.000** (0.000) | 0.000 (0.000) | 0.000 (0.000) | 0.000 (0.000) | 0.000 (0.000) |
| Education | -0.082 (0.051) | -0.042 (0.048) | -0.090* (0.047) | -0.019 (0.049) | -0.079* (0.047) |
| Observations | 859 | 859 | 859 | 859 | 859 |
| R-squared | 0.192 | 0.202 | 0.191 | 0.181 | 0.188 |
| F-statistic | 16.722 | 17.837 | 16.655 | 15.611 | 16.336 |
| ✱ Signif. codes: * p<0.10, ** p<0.05, *** p<0.01. Standard errors in parentheses. |
| Comb.Trans. | NoComb.Trans. | Mil.Aid | Fin.Aid | |
|---|---|---|---|---|
| Article9 | 0.021 (0.126) | -0.050 (0.123) | -0.008 (0.122) | -0.292** (0.122) |
| Democracy | -0.102 (0.116) | -0.165 (0.113) | -0.118 (0.113) | -0.138 (0.113) |
| DemocracyArticle9 | 0.169 (0.123) | 0.217* (0.120) | 0.206* (0.120) | -0.065 (0.119) |
| Partisanship | 0.172* (0.096) | 0.165* (0.094) | 0.147 (0.093) | 0.168* (0.093) |
| Ideology | -0.007 (0.021) | -0.005 (0.021) | -0.022 (0.021) | -0.004 (0.021) |
| Culture | 0.196*** (0.056) | 0.204*** (0.054) | 0.133** (0.054) | 0.098* (0.054) |
| Constitution | 0.837*** (0.089) | 0.782*** (0.087) | 0.533*** (0.086) | 0.172** (0.086) |
| Knowledge | -0.042 (0.120) | -0.060 (0.117) | -0.009 (0.116) | 0.050 (0.116) |
| Female | -0.469*** (0.091) | -0.496*** (0.088) | -0.136 (0.088) | -0.112 (0.088) |
| Age | -0.003 (0.003) | -0.001 (0.002) | -0.006** (0.002) | 0.002 (0.002) |
| Income | 0.000 (0.000) | 0.000 (0.000) | 0.000 (0.000) | 0.000 (0.000) |
| Education | 0.022 (0.048) | 0.022 (0.047) | -0.026 (0.047) | 0.041 (0.047) |
| Observations | 859 | 859 | 859 | 859 |
| R-squared | 0.192 | 0.195 | 0.092 | 0.031 |
| F-statistic | 16.704 | 17.061 | 7.147 | 2.279 |
| ✱ Signif. codes: * p<0.10, ** p<0.05, *** p<0.01. Standard errors in parentheses. |
B. Ordinary Least Squares Results by Manipulation-Check Status
Figure B1: Ordinary Least Squares results by manipulation-check status
data_ols <- security %>%
mutate(
Democracy = as.integer(group == "Democracy"),
Authoritarian = as.integer(group == "Authoritarian")
) %>%
select(
manipulation_pass,
starts_with("Q6.1_"),
Democracy, Authoritarian,
psu_rul, ideology, ML1, culture,
constitution, female, age, income, education
) %>%
drop_na()
fit_models <- function(df) {
models1 <- purrr::map(paste0("Q6.1_", 1:5), ~ lm(
reformulate(
c("Democracy","Authoritarian","psu_rul","ideology",
"ML1","culture","constitution","female","age","income","education"),
response = .x
), data = df
))
models2 <- purrr::map(paste0("Q6.1_", 6:9), ~ lm(
reformulate(
c("Democracy","Authoritarian","psu_rul","ideology",
"ML1","culture","constitution","female","age","income","education"),
response = .x
), data = df
))
list(models1 = models1, models2 = models2)
}
slices <- data_ols %>% split(.$manipulation_pass)
results <- purrr::map(slices, fit_models)
for (pass_val in names(results)) {
names(results[[pass_val]]$models1) <- paste0("Q6.1_", 1:5)
names(results[[pass_val]]$models2) <- paste0("Q6.1_", 6:9)
}
coef_df <- purrr::map_dfr(
names(results),
function(pass_val) {
pass_num <- as.integer(as.logical(pass_val))
models <- c(results[[pass_val]]$models1,
results[[pass_val]]$models2)
purrr::map_dfr(
models,
~ broom::tidy(.x),
.id = "outcome"
) %>%
dplyr::filter(term %in% c("Democracy", "Authoritarian")) %>%
dplyr::transmute(
outcome,
term,
pass = pass_num,
estimate,
se = std.error,
lower = estimate - 1.96*std.error,
upper = estimate + 1.96*std.error,
sig = p.value < 0.05
)
}
) %>%
dplyr::mutate(
term = factor(term, levels = c("Democracy","Authoritarian")),
outcome = factor(
outcome,
levels = paste0("Q6.1_",1:9),
labels = c(
"Military","Comb.Weapon","NoComb.Weapon","Comb.Ammun.","NoComb.Ammun.",
"Comb.Trans.","NoComb.Trans.","Mil.Aid","Fin.Aid"
)
)
)
pd <- position_dodge(width = 0.6)
ggplot(coef_df,
aes(x = estimate,
y = term,
color = factor(pass),
group = factor(pass),
size = sig,
linewidth = sig)) +
geom_vline(xintercept = 0, linetype = "dashed") +
geom_errorbarh(aes(xmin = lower, xmax = upper),
height = 0,
position = pd) +
geom_point(position = pd, shape = 16) +
facet_wrap(~ outcome, ncol = 3) +
scale_color_manual(
name = "Manipulation Check",
values = c("0" = "blue", "1" = "black"),
labels = c("Not passed", "Passed")
) +
scale_size_manual(
name = "Significance",
values = c("FALSE" = 0.5, "TRUE" = 2),
labels = c("n.s.", "* p < .05")
) +
scale_linewidth_manual(
name = "Significance",
values = c("FALSE" = 0.5, "TRUE" = 1.2)
) +
theme_bw(base_size = 12) +
theme(
axis.title.y = element_blank(),
strip.text = element_text(size = 10),
legend.position = "bottom"
) +
labs(x = "Estimated Treatment Effect")
Table B1: Ordinary Least Squares estimation results for the sample that DID NOT pass the manipulation check (Study 1)
Table B2: Ordinary Least Squares estimation results for the sample that DID NOT pass the manipulation check (Study 1)
Table B3: Ordinary Least Squares estimation results for the sample that PASSED the manipulation check (Study 1)
Table B4: Ordinary Least Squares estimation results for the sample that PASSED the manipulation check (Study 1)
| Military | Comb.Weapon | NoComb.Weapon | Comb.Ammun. | NoComb.Ammun. | |
|---|---|---|---|---|---|
| Democracy | 0.232* (0.119) | 0.337*** (0.126) | 0.371*** (0.124) | 0.304** (0.123) | 0.294** (0.125) |
| Authoritarian | 0.311*** (0.106) | 0.282** (0.113) | 0.293*** (0.111) | 0.275** (0.110) | 0.297*** (0.112) |
| Partisanship | 0.274*** (0.103) | 0.261** (0.109) | 0.313*** (0.107) | 0.298*** (0.106) | 0.346*** (0.108) |
| Ideology | -0.022 (0.022) | 0.011 (0.023) | 0.011 (0.023) | 0.015 (0.023) | -0.007 (0.023) |
| Patriotism | 0.114** (0.048) | 0.062 (0.051) | 0.080 (0.050) | 0.052 (0.050) | 0.071 (0.051) |
| Culture | 0.133** (0.059) | 0.192*** (0.062) | 0.193*** (0.062) | 0.156** (0.061) | 0.135** (0.062) |
| Constitution | 0.267*** (0.096) | 0.328*** (0.102) | 0.406*** (0.101) | 0.388*** (0.100) | 0.492*** (0.102) |
| Female | -0.190** (0.090) | -0.221** (0.095) | -0.183* (0.094) | -0.154* (0.093) | -0.168* (0.094) |
| Age | -0.011*** (0.003) | -0.010*** (0.003) | -0.010*** (0.003) | -0.010*** (0.003) | -0.008*** (0.003) |
| Income | 0.000 (0.000) | 0.000 (0.000) | 0.000 (0.000) | 0.000 (0.000) | 0.000 (0.000) |
| Education | -0.083* (0.048) | -0.062 (0.051) | -0.082* (0.050) | -0.098** (0.050) | -0.120** (0.050) |
| Observations | 597 | 597 | 597 | 597 | 597 |
| R-squared | 0.109 | 0.109 | 0.130 | 0.120 | 0.125 |
| F-statistic | 6.517 | 6.481 | 7.973 | 7.240 | 7.570 |
| ✱ Signif. codes: * p<0.10, ** p<0.05, *** p<0.01. Standard errors in parentheses. |
| Comb.Trans. | NoComb.Trans. | Mil.Aid | Fin.Aid | |
|---|---|---|---|---|
| Democracy | 0.460*** (0.127) | 0.226* (0.130) | 0.271** (0.128) | 0.033 (0.131) |
| Authoritarian | 0.406*** (0.113) | 0.252** (0.117) | 0.346*** (0.115) | 0.071 (0.117) |
| Partisanship | 0.329*** (0.109) | 0.287** (0.112) | 0.209* (0.110) | 0.296*** (0.113) |
| Ideology | 0.030 (0.023) | -0.003 (0.024) | -0.021 (0.024) | -0.057** (0.024) |
| Patriotism | 0.087* (0.051) | 0.044 (0.053) | 0.067 (0.052) | -0.099* (0.053) |
| Culture | 0.152** (0.063) | 0.116* (0.065) | 0.214*** (0.063) | 0.113* (0.065) |
| Constitution | 0.360*** (0.103) | 0.419*** (0.106) | 0.339*** (0.104) | 0.288*** (0.106) |
| Female | -0.274*** (0.096) | -0.223** (0.098) | -0.144 (0.097) | -0.061 (0.099) |
| Age | -0.011*** (0.003) | -0.006** (0.003) | -0.009*** (0.003) | 0.000 (0.003) |
| Income | 0.000 (0.000) | 0.000 (0.000) | 0.000 (0.000) | 0.000 (0.000) |
| Education | -0.069 (0.051) | -0.113** (0.052) | -0.162*** (0.051) | -0.119** (0.052) |
| Observations | 597 | 597 | 597 | 597 |
| R-squared | 0.137 | 0.099 | 0.108 | 0.073 |
| F-statistic | 8.443 | 5.854 | 6.456 | 4.172 |
| ✱ Signif. codes: * p<0.10, ** p<0.05, *** p<0.01. Standard errors in parentheses. |
| Military | Comb.Weapon | NoComb.Weapon | Comb.Ammun. | NoComb.Ammun. | |
|---|---|---|---|---|---|
| Democracy | -0.105 (0.140) | 0.086 (0.142) | 0.114 (0.147) | -0.081 (0.142) | 0.069 (0.145) |
| Authoritarian | -0.400*** (0.140) | -0.363** (0.142) | -0.407*** (0.147) | -0.544*** (0.142) | -0.408*** (0.145) |
| Partisanship | 0.085 (0.105) | 0.031 (0.106) | 0.069 (0.110) | 0.151 (0.106) | 0.082 (0.108) |
| Ideology | 0.035 (0.022) | 0.015 (0.022) | -0.007 (0.023) | -0.001 (0.023) | -0.013 (0.023) |
| Patriotism | 0.094* (0.049) | 0.093* (0.049) | 0.075 (0.051) | 0.053 (0.049) | 0.017 (0.050) |
| Culture | 0.217*** (0.061) | 0.282*** (0.062) | 0.225*** (0.064) | 0.217*** (0.062) | 0.222*** (0.063) |
| Constitution | 0.238*** (0.092) | 0.267*** (0.093) | 0.327*** (0.096) | 0.301*** (0.094) | 0.280*** (0.095) |
| Female | -0.146* (0.088) | -0.377*** (0.089) | -0.340*** (0.093) | -0.331*** (0.090) | -0.367*** (0.092) |
| Age | -0.009*** (0.003) | -0.010*** (0.003) | -0.009*** (0.003) | -0.010*** (0.003) | -0.010*** (0.003) |
| Income | 0.000 (0.000) | 0.000 (0.000) | 0.000 (0.000) | 0.000 (0.000) | 0.000 (0.000) |
| Education | -0.051 (0.049) | -0.075 (0.050) | -0.040 (0.051) | 0.004 (0.050) | 0.010 (0.051) |
| Observations | 610 | 610 | 610 | 610 | 610 |
| R-squared | 0.096 | 0.147 | 0.133 | 0.142 | 0.132 |
| F-statistic | 5.797 | 9.371 | 8.361 | 8.967 | 8.257 |
| ✱ Signif. codes: * p<0.10, ** p<0.05, *** p<0.01. Standard errors in parentheses. |
| Comb.Trans. | NoComb.Trans. | Mil.Aid | Fin.Aid | |
|---|---|---|---|---|
| Democracy | -0.168 (0.141) | 0.135 (0.148) | 0.271* (0.148) | 0.369** (0.160) |
| Authoritarian | -0.598*** (0.141) | -0.371** (0.148) | -0.246* (0.148) | -0.259 (0.160) |
| Partisanship | 0.097 (0.105) | 0.098 (0.111) | 0.107 (0.111) | 0.375*** (0.119) |
| Ideology | -0.009 (0.022) | -0.020 (0.023) | -0.036 (0.023) | -0.049* (0.025) |
| Patriotism | 0.005 (0.049) | 0.060 (0.051) | 0.062 (0.052) | -0.045 (0.056) |
| Culture | 0.205*** (0.061) | 0.244*** (0.064) | 0.213*** (0.064) | 0.183*** (0.069) |
| Constitution | 0.327*** (0.093) | 0.294*** (0.097) | 0.302*** (0.097) | 0.133 (0.105) |
| Female | -0.371*** (0.089) | -0.326*** (0.093) | -0.265*** (0.094) | -0.226** (0.101) |
| Age | -0.009*** (0.003) | -0.009*** (0.003) | -0.011*** (0.003) | -0.006** (0.003) |
| Income | 0.000 (0.000) | 0.000 (0.000) | -0.000 (0.000) | -0.000 (0.000) |
| Education | 0.006 (0.049) | 0.006 (0.052) | -0.014 (0.052) | 0.061 (0.056) |
| Observations | 610 | 610 | 610 | 610 |
| R-squared | 0.144 | 0.127 | 0.118 | 0.107 |
| F-statistic | 9.162 | 7.890 | 7.279 | 6.491 |
| ✱ Signif. codes: * p<0.10, ** p<0.05, *** p<0.01. Standard errors in parentheses. |
Figure B2: Odds-ration results by logistic regression estimation (DV: Manipulation check passed (1) or not (0))
logit_mod2 <- glm(
manipulation_pass ~
psu_rul + ideology + ML1 + culture +
constitution + female + age + income + education,
data = data_ols,
family = binomial(link = "logit")
)
coef_logit2 <- broom::tidy(logit_mod2) %>%
filter(term != "(Intercept)") %>%
mutate(
OR = exp(estimate),
lowerOR = exp(estimate - 1.96 * std.error),
upperOR = exp(estimate + 1.96 * std.error),
sig = p.value < 0.05,
term = recode(term,
psu_rul = "Partisanship",
ideology = "Ideology",
ML1 = "Patriotism",
culture = "Culture",
constitution = "Constitution",
female = "Female",
age = "Age",
income = "Income",
education = "Education")
) %>%
mutate(
term = factor(
term,
levels = c(
"Education",
"Income",
"Age",
"Female",
"Constitution",
"Culture",
"Patriotism",
"Ideology",
"Partisanship"
)
)
)
ggplot(coef_logit2, aes(x = OR, y = term)) +
geom_vline(xintercept = 1, linetype = "dashed") +
geom_errorbarh(aes(xmin = lowerOR, xmax = upperOR),
height = 0.2,
color = ifelse(coef_logit2$sig, "black", "grey60")) +
geom_point(aes(size = sig),
color = ifelse(coef_logit2$sig, "black", "grey60")) +
scale_size_manual(values = c("FALSE" = 1.5, "TRUE" = 3)) +
theme_bw(base_size = 12) +
theme(
axis.title.y = element_blank(),
legend.position = "none"
) +
labs(
x = "Odds Ratio (95% CI)",
title = ""
)