Studying Regime Preferences in Arab Countries

As a young Tunisian who was active during the Arab spring protests, I had first-hand experience with the views of people in Arab countries on democracy and other regime types. I believe that perceptions towards regime preferences in the Arab world are influenced by many factors such as religion, education, and culture, among others.

A few days ago, I stumbled upon the World Values Survey database and decided to see the extent to which people in four Arab countries support democracy vis-a-vis other alternatives using its last wave (2017-2020).

I used ordered logistic regression to study the impact of different IVs on three dependent variables:
1- Having a democratic political system.
2- Having an army rule.
3- Having a system governed by religious law in which there are no political parties or elections.

Then, I included multiple individual-level independent variables. I included the respondent’s age in z-scores, education in years, and several dummy variables that code whether the person is female, unemployed, and from a specific country.

Results:

The dots represent coefficient values and the whiskers are 95% intervals. The vertical line corresponds to the null hypothesis of 0; any estimate that doesn’t overlap with this line is interpreted as statistically significant.

Age seems to have a slightly positive impact on support for religious rule and a slightly negative one on support for democracy. Moreover, education has a surprisingly negative effect on support for democracy and positive effects on support for army rule and religious law, whereas women appear to be less supportive of religious rule. IVs representing the respondent’s countries also seem to have an impact on their regime preferences. Tunisians support religious rule the most and Egyptians support democracy the least. We also see a consistently negative impact on support for democracy across all four countries.

all =rbind(tn,jr,lb,eg)
all = all %>% select(Q260, Q262, B_COUNTRY, Q275, Q279, Q239P, Q238P, Q237P)
all = all %>% mutate(
  uid = seq(1, nrow(.)),
  tunisia = ifelse(B_COUNTRY == 788, 1, 0),
  egypt = ifelse(B_COUNTRY == 818, 1, 0),
  lebanon = ifelse(B_COUNTRY == 422, 1, 0),
  jordan = ifelse(B_COUNTRY == 400, 1, 0))

all = all %>% mutate(
  religiouslaw = recode(Q239P, "-5:-1=NA; 1=4; 2=3; 3=2; 4=1"),
  democracy = recode(Q238P, "-5:-1=NA; 1=4; 2=3; 3=2; 4=1"),
  army = recode(Q237P, "-5:-1=NA; 1=4; 2=3; 3=2; 4=1"))
all = all %>% mutate(
  female = ifelse(Q260 == 2, 1, 0),
  unemployed = recode(Q279, "-5:-1=NA; 1:6=0; 7=1; 8=0;"))
all = all %>% mutate(
  edu = recode(Q275, "-5:-1=NA; 0=1; 1=2; 2=3; 3=4; 4=5; 5=6; 6=7; 7=8; 8=9"))

all = all %>% mutate(
  z_age = arm::rescale(Q262))

polr(as.factor(religiouslaw) ~ female + z_age + unemployed + z_edu + tunisia + egypt + lebanon + jordan, data = all, Hess=TRUE) -> s1
ss1 = tidy(s1)
ss1 = ss1 %>% mutate(model = "Religious Law")

polr(as.factor(army) ~ female + z_age + unemployed + z_edu + tunisia + egypt + lebanon + jordan, data = all, Hess=TRUE) -> s2
ss2 = tidy(s2)
ss2 = ss2 %>% mutate(model = "Army Rule")

polr(as.factor(democracy) ~ female + z_age + unemployed + z_edu + tunisia + egypt + lebanon + jordan, data = all, Hess=TRUE) -> s3
ss3 = tidy(s3)
ss3 = ss3 %>% mutate(model = "Democracy")

alls = rbind(ss1,ss2,ss3) %>%
  relabel_predictors(z_age = "Age",
                     female = "Female",
                     z_edu = "Education",
                     unemployed = "Unemployed",
                     tunisia = "Tunisia",
                     jordan = "Jordan",
                     lebanon = "Lebanon",
                     egypt = "Egypt")
alls = alls %>% group_by(model) %>% filter(term != "(Intercept)	")

dwplot(alls, dot_args = list(size = 3, aes(colour = model, shape = model))) +
  geom_vline(xintercept = 0, colour = "grey50", linetype = 2)  +
  theme_minimal() +
  xlab("Coefficient Estimate") +
  scale_colour_manual(labels = c("Support for Army Rule",
                                 "Support for Democracy",
                                 "Support for Religious Rule"),
                      name="Models",
                      values = c("#69b3a2", "purple", "black")) +
  scale_shape_discrete(labels = c("Support for Army Rule",
                                  "Support for Democracy",
                                  "Support for Religious Rule"),
                       name="Models") + 
  theme(text = element_text(size = 20, family = "saira")) + 
  theme(legend.position = "bottom") +
  ggtitle("Support for Different Regime Types in Four Arab Countries")