wrangling
General
Wrangling
## wrangling initial
mint_11 <- read_sav("/Users/harshitraj/Desktop/R Work harshit/mint_round11.sav")
#code from round 11(1)
mint_11 <- mint_11 %>%
clean_names() %>%
mutate(p_party_new = case_when(p_party == 1 ~ 1,
p_party == 2 ~ 2,
!p_party %in% c(98, 99) ~ 3,
T ~ 4),
p_party_new = labelled(p_party_new, c("BJP" = 1,
"INC" = 2,
"Others" = 3,
"Non-Identifier" = 4)),
religion_in_new = case_when(religion_in == 4 ~ 1,
religion_in == 2 ~ 2,
religion_in == 3 ~ 3,
!religion_in %in% c(1, 99) ~ 4,
T ~ 5),
religion_in_new = labelled(religion_in_new, c("Hindu" = 1,
"Muslim" = 2,
"Christian" = 3,
"Others" = 4,
"Undisclosed/Irreligious" = 5)),
caste_in_new = case_when(caste_in == 1 ~ 1,
caste_in == 2 ~ 2,
caste_in == 3 ~ 3,
caste_in == 4 ~ 4,
T ~ 5),
caste_in_new = labelled(caste_in_new, c("General" = 1,
"OBC" = 2,
"SC" = 3,
"ST" = 4,
"Undisclosed/Not-Applicable" = 5)),
q_p_party_1_new = replace_na(q_p_party_1, 0),
q_p_party_2_new = replace_na(q_p_party_2, 0),
q_p_party_3_new = replace_na(q_p_party_3, 0),
q_p_party_4_new = replace_na(q_p_party_4, 0),
partisan_index = (q_p_party_1_new + q_p_party_2_new +
q_p_party_3_new + q_p_party_4_new)/4,
partisan_index_np = case_when(partisan_index!= 0 & partisan_index <= 2 ~ 4,
partisan_index > 2 & partisan_index <= 3 ~ 3,
partisan_index==0 ~ 1,
T ~ 2),
partisan_index_np = labelled(partisan_index_np, c("Strong Partisan" = 4,
"Moderate Partisan" = 3,
"Weak Partisan" = 2,
"No Partisan"= 1)),
partisan_index = case_when(partisan_index!= 0 & partisan_index <= 2 ~ 1,
partisan_index > 2 & partisan_index <= 3 ~ 2,
T ~ 3),
partisan_index = labelled(partisan_index, c("Strong Partisan" = 1,
"Moderate Partisan" = 2,
"Weak Partisan" = 3)),
age_cb_new = case_when(birthyear <= 1980 ~ 1,
birthyear >= 1997 ~ 3,
T ~ 2),
age_cb_new = labelled(age_cb_new, c("Pre-Millennials" = 1,
"Millennials" = 2,
"Post-Millennials" = 3)),
sec_in_new = case_when(new_sec_in == 12 ~ 1,
new_sec_in == 11 ~ 2,
T ~ 3),
sec_in_new = labelled(sec_in_new, c("High SEC" = 1,
"Middle SEC" = 2,
"Low SEC" = 3)),
p_alliance = case_when(p_party %in% c(2,4,5,6,7,8,9,14) ~ 1,
p_party %in% c(1,15) ~ 2,
T ~ 3),
p_alliance = labelled(p_alliance, c("I.N.D.I.A" = 1,
"NDA" = 2,
"Non-Allied" = 3)),
p_alliance_old = case_when(p_party %in% c(2,4,5,6,7,8,9,14) ~ 1,
p_party %in% c(1,15) ~ 2,
p_party %in% c(98,99) ~ 4,
T ~ 3),
p_alliance_old = labelled(p_alliance_old, c("I.N.D.I.A" = 1,
"NDA" = 2,
"Non-Allied" = 3,
"Non-Identifier" = 4)))
## joining a new distnce calculated csv
mint11_df <- read.csv("~/Desktop/R Work harshit/mint11_df.csv", header=TRUE)
mint11_df_1<-mint_11 %>% left_join(mint11_df, by="caseid")
## NEW REGRESSION analysis wrangling
glimpse(mint11_df_1$p_party_new)
## dbl+lbl [1:12544] 3, 1, 4, 1, 1, 1, 4, 1, 1, 1, 2, 3, 1, 3, 2, 1, 2, 3, 1,...
## @ labels: Named num [1:4] 1 2 3 4
## ..- attr(*, "names")= chr [1:4] "BJP" "INC" "Others" "Non-Identifier"
mint11_df_1 <- mint11_df_1 %>% clean_names() %>%
mutate(p_party_new_1 = case_when(p_party == 1 ~ 1,
p_party == 2 ~ 2,
T ~ 3),
p_party_new_1 = labelled(p_party_new_1, c("BJP" = 1,
"INC" = 2,
"Others" = 3
)),
religion_in_new = case_when(religion_in == 4 ~ 1,
religion_in == 2 ~ 2,
T ~ 3),
religion_in_new = labelled(religion_in_new, c("Hindu" = 1,
"Muslim" = 2,
"others" = 3
)),
caste_in_new = case_when(caste_in == 1 ~ 1,
caste_in == 2 ~ 2,
caste_in == 3 ~ 3,
caste_in == 4 ~ 4,
T ~ 5),
caste_in_new = labelled(caste_in_new, c("General" = 1,
"OBC" = 2,
"SC" = 3,
"ST" = 4,
"Undisclosed/Not-Applicable" = 5)),
partisan_index_1 = (q_p_party_1+ q_p_party_2 +
q_p_party_3 + q_p_party_4)/4,
age_cb_new = labelled(age_cb_new, c("Pre-Millennials" = 1,
"Millennials" = 2,
"Post-Millennials" = 3)),
sec_in_new = case_when(new_sec_in == 12 ~ 1,
new_sec_in == 11 ~ 2,
T ~ 3),
sec_in_new = labelled(sec_in_new, c("High SEC" = 1,
"Middle SEC" = 2,
"Low SEC" = 3)),
p_alliance = case_when(p_party %in% c(2,4,5,6,7,8,9,14) ~ 1,
p_party %in% c(1,15) ~ 2,
T ~ 3),
p_alliance = labelled(p_alliance, c("I.N.D.I.A" = 1,
"NDA" = 2,
"Non-Allied" = 3)),
p_alliance_old = case_when(p_party %in% c(2,4,5,6,7,8,9,14) ~ 1,
p_party %in% c(1,15) ~ 2,
p_party %in% c(98,99) ~ 4,
T ~ 3),
p_alliance_old = labelled(p_alliance_old, c("I.N.D.I.A" = 1,
"NDA" = 2,
"Non-Allied" = 3,
"Non-Identifier" = 4)),
geo_region_in_cb_new = case_when(geo_region_in_cb%in%c(1,4)~ 1,
T~2),
geo_region_in_cb_new=labelled(geo_region_in_cb_new, c("NW" = 1,
"SE" = 2)),
geo_citytier_in=case_when(geo_citytier_in==1~1,
TRUE~2),
geo_citytier_in=labelled(geo_citytier_in, c("Tier-1" = 1,
"Rest_Tiers"=2)),
bordering=case_when(geo_state_in_y%in%c("Rajasthan","Gujrat","Punjab","Jammu & Kashmir","Laddakh")~1,
TRUE~2),
hub_dist_1=hub_dist/1000,
hub_dist_2=log(hub_dist_1),
birthyear1=2024-birthyear)
Variables coding
Gender- Male and Female
Education
Not applicable: Illiterate- 1
Literate but no formal schooling/School up to 4 years- 2
Schooling between 5–9 years- 3
High School pass (SSC/HSC)- 4
Diploma or college certificate but not a graduate- 5
Graduate or Post Graduate General (B.A., M. A., B Com, BSC etc.)- 6
Graduate or Post Graduate Professional (MBA, MD, PhD etc.)-7
age
age= 2024-Birth Year
Caste
General = 1, OBC = 2, SC = 3, ST = 4, Undisclosed/Not-Applicable =
5
Party
BJP= 1 INC = 2 Others= 3
religion
Hindu = 1 Muslim = 2 others = 3
Distance
log(Distance/1000)
Partisan Index
Strong Partisan = 4, Moderate Partisan = 3,
Weak Partisan = 2
No Partisan= 1
wrangling
regression
# q9 israel hamas
## Israel Hamas blame wrangling
mint11_df_1<-mint11_df_1 %>% mutate(q9=ifelse(q9 ==2, 0, q9))
mint11_df_1<-mint11_df_1 %>% mutate(p_party_new_1 = as.factor(p_party_new_1)) %>%
mutate(geo_region_in_cb_new=as.factor(geo_region_in_cb_new)) %>%
mutate(religion_in_new=as.factor(religion_in_new)) %>%
mutate(geo_region_in_cb=as.factor(geo_region_in_cb)) %>%
mutate(geo_citytier_in=as.factor(geo_citytier_in))
mint11_df_1 <- mint11_df_1 %>%
mutate(p_party_new_1 = fct_relevel(p_party_new_1, "2", after = 2),
p_party_new_1 = fct_relevel(p_party_new_1, "3", after = 0),
religion_in_new= fct_relevel(religion_in_new, "3", after = 0),
religion_in_new=fct_relevel(religion_in_new, "2", after = 2))
Support For
Nuclear
#q10 regression
## q10 wrangling
mint11_df_1<-mint11_df_1 %>% mutate(q10=ifelse(q10 ==2, 0, q10))
## new_q10_1 DV <- age + partisanship + party.
base_new_model_q10_1 <-mint11_df_1 %>% glm(q10 ~ partisan_index_np+birthyear1+p_party_new_1+religion_in_new+hub_dist_2, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
vif(base_new_model_q10_1)
## GVIF Df GVIF^(1/(2*Df))
## partisan_index_np 1.341150 1 1.158080
## birthyear1 1.008124 1 1.004054
## p_party_new_1 1.599940 2 1.124672
## religion_in_new 1.216770 2 1.050273
## hub_dist_2 1.028003 1 1.013905
names(base_new_model_q10_1$coefficients) <- c("intercept","Partisan_Index","age","BJP","INC","Hindu","Muslim","distance")
summary(base_new_model_q10_1)
##
## Call:
## glm(formula = q10 ~ partisan_index_np + birthyear1 + p_party_new_1 +
## religion_in_new + hub_dist_2, family = binomial, data = .,
## weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept -1.475693 0.091866 -16.063 < 2e-16 ***
## Partisan_Index 0.339867 0.021286 15.966 < 2e-16 ***
## age -0.010509 0.001851 -5.679 1.36e-08 ***
## BJP 0.140532 0.052366 2.684 0.00728 **
## INC -0.155617 0.075516 -2.061 0.03933 *
## Hindu 0.361696 0.057624 6.277 3.46e-10 ***
## Muslim -0.346405 0.083962 -4.126 3.70e-05 ***
## distance -0.105507 0.026892 -3.923 8.73e-05 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 14700 on 11161 degrees of freedom
## Residual deviance: 13968 on 11154 degrees of freedom
## (1382 observations deleted due to missingness)
## AIC: 13888
##
## Number of Fisher Scoring iterations: 4
summ(model =base_new_model_q10_1, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
11162 (1382 missing obs. deleted)
|
|
Dependent variable
|
q10
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(7)
|
732.20
|
|
Pseudo-R² (Cragg-Uhler)
|
0.09
|
|
Pseudo-R² (McFadden)
|
0.05
|
|
AIC
|
13887.99
|
|
BIC
|
13946.55
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.23
|
0.19
|
0.27
|
-16.06
|
0.00
|
|
Partisan_Index
|
1.40
|
1.35
|
1.46
|
15.97
|
0.00
|
|
age
|
0.99
|
0.99
|
0.99
|
-5.68
|
0.00
|
|
BJP
|
1.15
|
1.04
|
1.28
|
2.68
|
0.01
|
|
INC
|
0.86
|
0.74
|
0.99
|
-2.06
|
0.04
|
|
Hindu
|
1.44
|
1.28
|
1.61
|
6.28
|
0.00
|
|
Muslim
|
0.71
|
0.60
|
0.83
|
-4.13
|
0.00
|
|
distance
|
0.90
|
0.85
|
0.95
|
-3.92
|
0.00
|
|
Standard errors: MLE
|
##new model 10_2
base_new_model_q10_2 <-mint11_df_1 %>% glm(q10 ~ partisan_index_np+birthyear1+p_party_new_1+religion_in_new+hub_dist_2+education_india, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
vif(base_new_model_q10_2)
## GVIF Df GVIF^(1/(2*Df))
## partisan_index_np 1.342000 1 1.158447
## birthyear1 1.034105 1 1.016909
## p_party_new_1 1.602290 2 1.125085
## religion_in_new 1.222759 2 1.051563
## hub_dist_2 1.028067 1 1.013936
## education_india 1.035664 1 1.017676
names(base_new_model_q10_2$coefficients) <- c("intercept","Partisan_Index","age","BJP","INC","Hindu","Muslim","distance","Education")
summary(base_new_model_q10_2)
##
## Call:
## glm(formula = q10 ~ partisan_index_np + birthyear1 + p_party_new_1 +
## religion_in_new + hub_dist_2 + education_india, family = binomial,
## data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept -1.313525 0.144174 -9.111 < 2e-16 ***
## Partisan_Index 0.338990 0.021295 15.919 < 2e-16 ***
## age -0.010076 0.001873 -5.381 7.42e-08 ***
## BJP 0.142998 0.052401 2.729 0.00635 **
## INC -0.152360 0.075551 -2.017 0.04373 *
## Hindu 0.364590 0.057667 6.322 2.58e-10 ***
## Muslim -0.350552 0.084015 -4.173 3.01e-05 ***
## distance -0.105790 0.026894 -3.934 8.37e-05 ***
## Education -0.029914 0.020538 -1.456 0.14525
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 14700 on 11161 degrees of freedom
## Residual deviance: 13965 on 11153 degrees of freedom
## (1382 observations deleted due to missingness)
## AIC: 13888
##
## Number of Fisher Scoring iterations: 4
summ(model =base_new_model_q10_2, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
11162 (1382 missing obs. deleted)
|
|
Dependent variable
|
q10
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(8)
|
734.32
|
|
Pseudo-R² (Cragg-Uhler)
|
0.09
|
|
Pseudo-R² (McFadden)
|
0.05
|
|
AIC
|
13888.38
|
|
BIC
|
13954.27
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.27
|
0.20
|
0.36
|
-9.11
|
0.00
|
|
Partisan_Index
|
1.40
|
1.35
|
1.46
|
15.92
|
0.00
|
|
age
|
0.99
|
0.99
|
0.99
|
-5.38
|
0.00
|
|
BJP
|
1.15
|
1.04
|
1.28
|
2.73
|
0.01
|
|
INC
|
0.86
|
0.74
|
1.00
|
-2.02
|
0.04
|
|
Hindu
|
1.44
|
1.29
|
1.61
|
6.32
|
0.00
|
|
Muslim
|
0.70
|
0.60
|
0.83
|
-4.17
|
0.00
|
|
distance
|
0.90
|
0.85
|
0.95
|
-3.93
|
0.00
|
|
Education
|
0.97
|
0.93
|
1.01
|
-1.46
|
0.15
|
|
Standard errors: MLE
|
## new model 10_3
base_new_model_q10_3 <-mint11_df_1 %>% glm(q10 ~ partisan_index_np+birthyear1+p_party_new_1+religion_in_new+hub_dist_2+education_india+caste_in_new, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
vif(base_new_model_q10_3)
## GVIF Df GVIF^(1/(2*Df))
## partisan_index_np 1.343099 1 1.158921
## birthyear1 1.048806 1 1.024112
## p_party_new_1 1.616956 2 1.127651
## religion_in_new 1.242948 2 1.055877
## hub_dist_2 1.037236 1 1.018448
## education_india 1.039539 1 1.019578
## caste_in_new 1.082535 1 1.040450
names(base_new_model_q10_3$coefficients) <- c("intercept","Partisan_Index","age","BJP","INC","Hindu","Muslim","distance","Education","Caste")
summary(base_new_model_q10_3)
##
## Call:
## glm(formula = q10 ~ partisan_index_np + birthyear1 + p_party_new_1 +
## religion_in_new + hub_dist_2 + education_india + caste_in_new,
## family = binomial, data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept -1.253656 0.156082 -8.032 9.58e-16 ***
## Partisan_Index 0.338236 0.021306 15.875 < 2e-16 ***
## age -0.010298 0.001886 -5.460 4.75e-08 ***
## BJP 0.138027 0.052634 2.622 0.008731 **
## INC -0.153932 0.075561 -2.037 0.041630 *
## Hindu 0.357456 0.058107 6.152 7.67e-10 ***
## Muslim -0.358665 0.084412 -4.249 2.15e-05 ***
## distance -0.103246 0.027019 -3.821 0.000133 ***
## Education -0.031156 0.020578 -1.514 0.130011
## Caste -0.018855 0.018896 -0.998 0.318357
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 14700 on 11161 degrees of freedom
## Residual deviance: 13964 on 11152 degrees of freedom
## (1382 observations deleted due to missingness)
## AIC: 13889
##
## Number of Fisher Scoring iterations: 4
summ(model =base_new_model_q10_3, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
11162 (1382 missing obs. deleted)
|
|
Dependent variable
|
q10
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(9)
|
735.32
|
|
Pseudo-R² (Cragg-Uhler)
|
0.09
|
|
Pseudo-R² (McFadden)
|
0.05
|
|
AIC
|
13889.49
|
|
BIC
|
13962.70
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.29
|
0.21
|
0.39
|
-8.03
|
0.00
|
|
Partisan_Index
|
1.40
|
1.35
|
1.46
|
15.87
|
0.00
|
|
age
|
0.99
|
0.99
|
0.99
|
-5.46
|
0.00
|
|
BJP
|
1.15
|
1.04
|
1.27
|
2.62
|
0.01
|
|
INC
|
0.86
|
0.74
|
0.99
|
-2.04
|
0.04
|
|
Hindu
|
1.43
|
1.28
|
1.60
|
6.15
|
0.00
|
|
Muslim
|
0.70
|
0.59
|
0.82
|
-4.25
|
0.00
|
|
distance
|
0.90
|
0.86
|
0.95
|
-3.82
|
0.00
|
|
Education
|
0.97
|
0.93
|
1.01
|
-1.51
|
0.13
|
|
Caste
|
0.98
|
0.95
|
1.02
|
-1.00
|
0.32
|
|
Standard errors: MLE
|
## new model 10_4
base_new_model_q10_4 <-mint11_df_1 %>% glm(q10 ~ partisan_index_np+birthyear1+p_party_new_1+religion_in_new+hub_dist_2+education_india+caste_in_new+gender, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
vif(base_new_model_q10_4)
## GVIF Df GVIF^(1/(2*Df))
## partisan_index_np 1.347284 1 1.160725
## birthyear1 1.049081 1 1.024247
## p_party_new_1 1.619922 2 1.128167
## religion_in_new 1.243693 2 1.056035
## hub_dist_2 1.037261 1 1.018460
## education_india 1.043523 1 1.021530
## caste_in_new 1.083498 1 1.040912
## gender 1.014136 1 1.007043
names(base_new_model_q10_4$coefficients) <- c("intercept","Partisan_Index","age","BJP","INC","Hindu","Muslim","distance","Education","Caste","gender")
summary(base_new_model_q10_4)
##
## Call:
## glm(formula = q10 ~ partisan_index_np + birthyear1 + p_party_new_1 +
## religion_in_new + hub_dist_2 + education_india + caste_in_new +
## gender, family = binomial, data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept -1.052233 0.167260 -6.291 3.15e-10 ***
## Partisan_Index 0.334023 0.021347 15.647 < 2e-16 ***
## age -0.010339 0.001885 -5.486 4.12e-08 ***
## BJP 0.132093 0.052688 2.507 0.012174 *
## INC -0.150500 0.075636 -1.990 0.046614 *
## Hindu 0.352557 0.058149 6.063 1.34e-09 ***
## Muslim -0.364342 0.084481 -4.313 1.61e-05 ***
## distance -0.101626 0.027029 -3.760 0.000170 ***
## Education -0.026986 0.020624 -1.308 0.190710
## Caste -0.020746 0.018921 -1.096 0.272877
## gender -0.136311 0.040853 -3.337 0.000848 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 14700 on 11161 degrees of freedom
## Residual deviance: 13953 on 11151 degrees of freedom
## (1382 observations deleted due to missingness)
## AIC: 13880
##
## Number of Fisher Scoring iterations: 4
summ(model =base_new_model_q10_4, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
11162 (1382 missing obs. deleted)
|
|
Dependent variable
|
q10
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(10)
|
746.45
|
|
Pseudo-R² (Cragg-Uhler)
|
0.09
|
|
Pseudo-R² (McFadden)
|
0.05
|
|
AIC
|
13879.68
|
|
BIC
|
13960.21
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.35
|
0.25
|
0.48
|
-6.29
|
0.00
|
|
Partisan_Index
|
1.40
|
1.34
|
1.46
|
15.65
|
0.00
|
|
age
|
0.99
|
0.99
|
0.99
|
-5.49
|
0.00
|
|
BJP
|
1.14
|
1.03
|
1.27
|
2.51
|
0.01
|
|
INC
|
0.86
|
0.74
|
1.00
|
-1.99
|
0.05
|
|
Hindu
|
1.42
|
1.27
|
1.59
|
6.06
|
0.00
|
|
Muslim
|
0.69
|
0.59
|
0.82
|
-4.31
|
0.00
|
|
distance
|
0.90
|
0.86
|
0.95
|
-3.76
|
0.00
|
|
Education
|
0.97
|
0.93
|
1.01
|
-1.31
|
0.19
|
|
Caste
|
0.98
|
0.94
|
1.02
|
-1.10
|
0.27
|
|
gender
|
0.87
|
0.81
|
0.95
|
-3.34
|
0.00
|
|
Standard errors: MLE
|
Support For Nuclear
Combined
base_combined_new10_1<- stargazer(base_new_model_q10_1,base_new_model_q10_2,base_new_model_q10_3,base_new_model_q10_4, title = "Results: Support For Use Of Nuclear", align = TRUE, type = "text",
dep.var.caption = "Dependent Variable: Support For Use Of Nuclear",column.labels = c("Model 1", "Model 2","Model 3","Model 4"),
dep.var.labels.include = FALSE,
font.size = "normalsize")
##
## Results: Support For Use Of Nuclear
## ===================================================================
## Dependent Variable: Support For Use Of Nuclear
## -------------------------------------------------
## Model 1 Model 2 Model 3 Model 4
## (1) (2) (3) (4)
## -------------------------------------------------------------------
## intercept -1.476*** -1.314*** -1.254*** -1.052***
## (0.092) (0.144) (0.156) (0.167)
##
## Partisan_Index 0.340*** 0.339*** 0.338*** 0.334***
## (0.021) (0.021) (0.021) (0.021)
##
## age -0.011*** -0.010*** -0.010*** -0.010***
## (0.002) (0.002) (0.002) (0.002)
##
## BJP 0.141*** 0.143*** 0.138*** 0.132**
## (0.052) (0.052) (0.053) (0.053)
##
## INC -0.156** -0.152** -0.154** -0.150**
## (0.076) (0.076) (0.076) (0.076)
##
## Hindu 0.362*** 0.365*** 0.357*** 0.353***
## (0.058) (0.058) (0.058) (0.058)
##
## Muslim -0.346*** -0.351*** -0.359*** -0.364***
## (0.084) (0.084) (0.084) (0.084)
##
## distance -0.106*** -0.106*** -0.103*** -0.102***
## (0.027) (0.027) (0.027) (0.027)
##
## Education -0.030 -0.031 -0.027
## (0.021) (0.021) (0.021)
##
## Caste -0.019 -0.021
## (0.019) (0.019)
##
## gender -0.136***
## (0.041)
##
## -------------------------------------------------------------------
## Observations 11,162 11,162 11,162 11,162
## Log Likelihood -6,935.993 -6,935.192 -6,934.746 -6,928.842
## Akaike Inf. Crit. 13,887.990 13,888.380 13,889.490 13,879.680
## ===================================================================
## Note: *p<0.1; **p<0.05; ***p<0.01
Support For Nuclear if
50% Chances
# question 11a
mint11_df_9<-mint11_df_1 %>%filter(!is.na(q11a)) %>%mutate(q11a_1=ifelse(q11a ==2, 0, q11a))
## new_q11a_1
base_new_model_q11a_1 <-mint11_df_9 %>% glm(q11a_1 ~ partisan_index_np+birthyear1+p_party_new_1+religion_in_new+hub_dist_2, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
vif(base_new_model_q11a_1)
## GVIF Df GVIF^(1/(2*Df))
## partisan_index_np 1.374288 1 1.172300
## birthyear1 1.001477 1 1.000738
## p_party_new_1 1.629914 2 1.129903
## religion_in_new 1.208302 2 1.048441
## hub_dist_2 1.023008 1 1.011438
names(base_new_model_q11a_1$coefficients) <- c("intercept","Partisan_Index","age","BJP","INC","Hindu","Muslim","distance")
summary(base_new_model_q11a_1)
##
## Call:
## glm(formula = q11a_1 ~ partisan_index_np + birthyear1 + p_party_new_1 +
## religion_in_new + hub_dist_2, family = binomial, data = .,
## weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept -0.565754 0.159381 -3.550 0.000386 ***
## Partisan_Index 0.205801 0.035174 5.851 4.89e-09 ***
## age 0.018397 0.003250 5.661 1.51e-08 ***
## BJP 0.063397 0.086191 0.736 0.462005
## INC 0.093252 0.131101 0.711 0.476896
## Hindu -0.075727 0.098807 -0.766 0.443431
## Muslim -0.446862 0.147565 -3.028 0.002460 **
## distance -0.009121 0.045550 -0.200 0.841284
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5438.5 on 4162 degrees of freedom
## Residual deviance: 5338.3 on 4155 degrees of freedom
## (418 observations deleted due to missingness)
## AIC: 5220.6
##
## Number of Fisher Scoring iterations: 4
summ(model = base_new_model_q11a_1, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
4163 (418 missing obs. deleted)
|
|
Dependent variable
|
q11a_1
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(7)
|
100.18
|
|
Pseudo-R² (Cragg-Uhler)
|
0.03
|
|
Pseudo-R² (McFadden)
|
0.02
|
|
AIC
|
5220.57
|
|
BIC
|
5271.24
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.57
|
0.42
|
0.78
|
-3.55
|
0.00
|
|
Partisan_Index
|
1.23
|
1.15
|
1.32
|
5.85
|
0.00
|
|
age
|
1.02
|
1.01
|
1.03
|
5.66
|
0.00
|
|
BJP
|
1.07
|
0.90
|
1.26
|
0.74
|
0.46
|
|
INC
|
1.10
|
0.85
|
1.42
|
0.71
|
0.48
|
|
Hindu
|
0.93
|
0.76
|
1.13
|
-0.77
|
0.44
|
|
Muslim
|
0.64
|
0.48
|
0.85
|
-3.03
|
0.00
|
|
distance
|
0.99
|
0.91
|
1.08
|
-0.20
|
0.84
|
|
Standard errors: MLE
|
## new_q11a_2
base_new_model_q11a_2 <-mint11_df_9 %>% glm(q11a_1 ~ partisan_index_np+birthyear1+p_party_new_1+religion_in_new+hub_dist_2+education_india, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
vif(base_new_model_q11a_2)
## GVIF Df GVIF^(1/(2*Df))
## partisan_index_np 1.375932 1 1.173001
## birthyear1 1.029277 1 1.014533
## p_party_new_1 1.636593 2 1.131059
## religion_in_new 1.211286 2 1.049087
## hub_dist_2 1.022923 1 1.011396
## education_india 1.036924 1 1.018295
names(base_new_model_q11a_2$coefficients) <- c("intercept","Partisan_Index","age","BJP","INC","Hindu","Muslim","distance","Education")
summary(base_new_model_q11a_2)
##
## Call:
## glm(formula = q11a_1 ~ partisan_index_np + birthyear1 + p_party_new_1 +
## religion_in_new + hub_dist_2 + education_india, family = binomial,
## data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept -0.720805 0.234998 -3.067 0.00216 **
## Partisan_Index 0.206917 0.035201 5.878 4.15e-09 ***
## age 0.017909 0.003290 5.444 5.21e-08 ***
## BJP 0.058697 0.086358 0.680 0.49670
## INC 0.088266 0.131226 0.673 0.50119
## Hindu -0.077988 0.098853 -0.789 0.43016
## Muslim -0.443829 0.147633 -3.006 0.00264 **
## distance -0.009173 0.045532 -0.201 0.84033
## Education 0.029122 0.032424 0.898 0.36910
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5438.5 on 4162 degrees of freedom
## Residual deviance: 5337.5 on 4154 degrees of freedom
## (418 observations deleted due to missingness)
## AIC: 5221.5
##
## Number of Fisher Scoring iterations: 4
summ(model =base_new_model_q11a_2, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
4163 (418 missing obs. deleted)
|
|
Dependent variable
|
q11a_1
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(8)
|
100.98
|
|
Pseudo-R² (Cragg-Uhler)
|
0.03
|
|
Pseudo-R² (McFadden)
|
0.02
|
|
AIC
|
5221.50
|
|
BIC
|
5278.51
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.49
|
0.31
|
0.77
|
-3.07
|
0.00
|
|
Partisan_Index
|
1.23
|
1.15
|
1.32
|
5.88
|
0.00
|
|
age
|
1.02
|
1.01
|
1.02
|
5.44
|
0.00
|
|
BJP
|
1.06
|
0.90
|
1.26
|
0.68
|
0.50
|
|
INC
|
1.09
|
0.84
|
1.41
|
0.67
|
0.50
|
|
Hindu
|
0.92
|
0.76
|
1.12
|
-0.79
|
0.43
|
|
Muslim
|
0.64
|
0.48
|
0.86
|
-3.01
|
0.00
|
|
distance
|
0.99
|
0.91
|
1.08
|
-0.20
|
0.84
|
|
Education
|
1.03
|
0.97
|
1.10
|
0.90
|
0.37
|
|
Standard errors: MLE
|
# new_q11a_3
base_new_model_q11a_3 <-mint11_df_9 %>% glm(q11a_1 ~ partisan_index_np+birthyear1+p_party_new_1+religion_in_new+hub_dist_1+education_india+caste_in_new, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
vif(base_new_model_q11a_3)
## GVIF Df GVIF^(1/(2*Df))
## partisan_index_np 1.377576 1 1.173702
## birthyear1 1.041200 1 1.020392
## p_party_new_1 1.658080 2 1.134753
## religion_in_new 1.229537 2 1.053017
## hub_dist_1 1.043088 1 1.021317
## education_india 1.042537 1 1.021047
## caste_in_new 1.067362 1 1.033132
names(base_new_model_q11a_3$coefficients) <- c("intercept","Partisan_Index","age","BJP","INC","Hindu","Muslim","distance","Education","Caste")
summary(base_new_model_q11a_3)
##
## Call:
## glm(formula = q11a_1 ~ partisan_index_np + birthyear1 + p_party_new_1 +
## religion_in_new + hub_dist_1 + education_india + caste_in_new,
## family = binomial, data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept -0.61088 0.26055 -2.345 0.01905 *
## Partisan_Index 0.20701 0.03523 5.876 4.20e-09 ***
## age 0.01780 0.00331 5.378 7.53e-08 ***
## BJP 0.04570 0.08684 0.526 0.59870
## INC 0.08889 0.13125 0.677 0.49823
## Hindu -0.08548 0.09961 -0.858 0.39077
## Muslim -0.44937 0.14824 -3.031 0.00243 **
## distance -0.07046 0.06309 -1.117 0.26406
## Education 0.02757 0.03253 0.847 0.39675
## Caste -0.01058 0.03104 -0.341 0.73315
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5438.5 on 4162 degrees of freedom
## Residual deviance: 5336.1 on 4153 degrees of freedom
## (418 observations deleted due to missingness)
## AIC: 5221.1
##
## Number of Fisher Scoring iterations: 4
summ(model =base_new_model_q11a_3, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
4163 (418 missing obs. deleted)
|
|
Dependent variable
|
q11a_1
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(9)
|
102.39
|
|
Pseudo-R² (Cragg-Uhler)
|
0.03
|
|
Pseudo-R² (McFadden)
|
0.02
|
|
AIC
|
5221.05
|
|
BIC
|
5284.39
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.54
|
0.33
|
0.90
|
-2.34
|
0.02
|
|
Partisan_Index
|
1.23
|
1.15
|
1.32
|
5.88
|
0.00
|
|
age
|
1.02
|
1.01
|
1.02
|
5.38
|
0.00
|
|
BJP
|
1.05
|
0.88
|
1.24
|
0.53
|
0.60
|
|
INC
|
1.09
|
0.85
|
1.41
|
0.68
|
0.50
|
|
Hindu
|
0.92
|
0.76
|
1.12
|
-0.86
|
0.39
|
|
Muslim
|
0.64
|
0.48
|
0.85
|
-3.03
|
0.00
|
|
distance
|
0.93
|
0.82
|
1.05
|
-1.12
|
0.26
|
|
Education
|
1.03
|
0.96
|
1.10
|
0.85
|
0.40
|
|
Caste
|
0.99
|
0.93
|
1.05
|
-0.34
|
0.73
|
|
Standard errors: MLE
|
## new model q11a_4
base_new_model_q11a_4 <-mint11_df_9 %>% glm(q11a_1 ~ partisan_index_np+birthyear1+p_party_new_1+religion_in_new+hub_dist_2+education_india+caste_in_new+gender, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
vif(base_new_model_q11a_4)
## GVIF Df GVIF^(1/(2*Df))
## partisan_index_np 1.378173 1 1.173956
## birthyear1 1.040930 1 1.020260
## p_party_new_1 1.645548 2 1.132603
## religion_in_new 1.229198 2 1.052944
## hub_dist_2 1.032506 1 1.016123
## education_india 1.045576 1 1.022534
## caste_in_new 1.068629 1 1.033745
## gender 1.013159 1 1.006558
names(base_new_model_q11a_4$coefficients) <- c("intercept","Partisan_Index","age","BJP","INC","Hindu","Muslim","distance","Education","Caste","Gender")
summary(base_new_model_q11a_4)
##
## Call:
## glm(formula = q11a_1 ~ partisan_index_np + birthyear1 + p_party_new_1 +
## religion_in_new + hub_dist_2 + education_india + caste_in_new +
## gender, family = binomial, data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept -0.269394 0.270007 -0.998 0.31841
## Partisan_Index 0.197421 0.035341 5.586 2.32e-08 ***
## age 0.018198 0.003328 5.468 4.54e-08 ***
## BJP 0.051817 0.086712 0.598 0.55013
## INC 0.111077 0.131645 0.844 0.39880
## Hindu -0.081417 0.099831 -0.816 0.41476
## Muslim -0.446319 0.148725 -3.001 0.00269 **
## distance -0.004751 0.045838 -0.104 0.91745
## Education 0.037491 0.032611 1.150 0.25030
## Caste -0.016593 0.031100 -0.534 0.59366
## Gender -0.304207 0.066240 -4.592 4.38e-06 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5438.5 on 4162 degrees of freedom
## Residual deviance: 5316.2 on 4152 degrees of freedom
## (418 observations deleted due to missingness)
## AIC: 5203.1
##
## Number of Fisher Scoring iterations: 4
summ(model =base_new_model_q11a_4, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
4163 (418 missing obs. deleted)
|
|
Dependent variable
|
q11a_1
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(10)
|
122.27
|
|
Pseudo-R² (Cragg-Uhler)
|
0.04
|
|
Pseudo-R² (McFadden)
|
0.02
|
|
AIC
|
5203.06
|
|
BIC
|
5272.74
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.76
|
0.45
|
1.30
|
-1.00
|
0.32
|
|
Partisan_Index
|
1.22
|
1.14
|
1.31
|
5.59
|
0.00
|
|
age
|
1.02
|
1.01
|
1.03
|
5.47
|
0.00
|
|
BJP
|
1.05
|
0.89
|
1.25
|
0.60
|
0.55
|
|
INC
|
1.12
|
0.86
|
1.45
|
0.84
|
0.40
|
|
Hindu
|
0.92
|
0.76
|
1.12
|
-0.82
|
0.41
|
|
Muslim
|
0.64
|
0.48
|
0.86
|
-3.00
|
0.00
|
|
distance
|
1.00
|
0.91
|
1.09
|
-0.10
|
0.92
|
|
Education
|
1.04
|
0.97
|
1.11
|
1.15
|
0.25
|
|
Caste
|
0.98
|
0.93
|
1.05
|
-0.53
|
0.59
|
|
Gender
|
0.74
|
0.65
|
0.84
|
-4.59
|
0.00
|
|
Standard errors: MLE
|
Combined If 50 %
base_combined_new11_1<- stargazer(base_new_model_q11a_1,base_new_model_q11a_2,base_new_model_q11a_3,base_new_model_q11a_4, title = "Results: Support For Use Of Nuclear If 50%", align = TRUE, type = "text",
dep.var.caption = "Dependent Variable: Support For Use Of Nuclear If 50%",column.labels = c("Model 1", "Model 2","Model 3","Model 4"),
dep.var.labels.include = FALSE,
font.size = "normalsize")
##
## Results: Support For Use Of Nuclear If 50%
## ==========================================================================
## Dependent Variable: Support For Use Of Nuclear If 50%
## --------------------------------------------------------
## Model 1 Model 2 Model 3 Model 4
## (1) (2) (3) (4)
## --------------------------------------------------------------------------
## intercept -0.566*** -0.721*** -0.611** -0.269
## (0.159) (0.235) (0.261) (0.270)
##
## Partisan_Index 0.206*** 0.207*** 0.207*** 0.197***
## (0.035) (0.035) (0.035) (0.035)
##
## age 0.018*** 0.018*** 0.018*** 0.018***
## (0.003) (0.003) (0.003) (0.003)
##
## BJP 0.063 0.059 0.046 0.052
## (0.086) (0.086) (0.087) (0.087)
##
## INC 0.093 0.088 0.089 0.111
## (0.131) (0.131) (0.131) (0.132)
##
## Hindu -0.076 -0.078 -0.085 -0.081
## (0.099) (0.099) (0.100) (0.100)
##
## Muslim -0.447*** -0.444*** -0.449*** -0.446***
## (0.148) (0.148) (0.148) (0.149)
##
## distance -0.009 -0.009 -0.070 -0.005
## (0.046) (0.046) (0.063) (0.046)
##
## Education 0.029 0.028 0.037
## (0.032) (0.033) (0.033)
##
## Caste -0.011 -0.017
## (0.031) (0.031)
##
## Gender -0.304***
## (0.066)
##
## --------------------------------------------------------------------------
## Observations 4,163 4,163 4,163 4,163
## Log Likelihood -2,602.285 -2,601.751 -2,600.527 -2,590.532
## Akaike Inf. Crit. 5,220.570 5,221.502 5,221.054 5,203.064
## ==========================================================================
## Note: *p<0.1; **p<0.05; ***p<0.01