# library and wrangling
library(tidyverse)
## ── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
## ✔ dplyr 1.1.4 ✔ readr 2.1.5
## ✔ forcats 1.0.0 ✔ stringr 1.5.1
## ✔ ggplot2 3.4.4 ✔ tibble 3.2.1
## ✔ lubridate 1.9.3 ✔ tidyr 1.3.1
## ✔ purrr 1.0.2
## ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
## ✖ dplyr::filter() masks stats::filter()
## ✖ dplyr::lag() masks stats::lag()
## ℹ Use the conflicted package (<http://conflicted.r-lib.org/>) to force all conflicts to become errors
library(haven)
library(labelled)
library(janitor)
##
## Attaching package: 'janitor'
##
## The following objects are masked from 'package:stats':
##
## chisq.test, fisher.test
library(openxlsx)
library(kableExtra)
##
## Attaching package: 'kableExtra'
##
## The following object is masked from 'package:dplyr':
##
## group_rows
library(forcats)
library(Hmisc)
##
## Attaching package: 'Hmisc'
##
## The following objects are masked from 'package:dplyr':
##
## src, summarize
##
## The following objects are masked from 'package:base':
##
## format.pval, units
library(jtools)
##
## Attaching package: 'jtools'
##
## The following object is masked from 'package:Hmisc':
##
## %nin%
library(sjstats)
## Registered S3 methods overwritten by 'broom':
## method from
## tidy.glht jtools
## tidy.summary.glht jtools
library(arm)
## Loading required package: MASS
##
## Attaching package: 'MASS'
##
## The following object is masked from 'package:dplyr':
##
## select
##
## Loading required package: Matrix
##
## Attaching package: 'Matrix'
##
## The following objects are masked from 'package:tidyr':
##
## expand, pack, unpack
##
## Loading required package: lme4
##
## arm (Version 1.14-4, built: 2024-4-1)
##
## Working directory is /Users/harshitraj/Desktop/R Work harshit
##
##
## Attaching package: 'arm'
##
## The following object is masked from 'package:jtools':
##
## standardize
library(car)
## Loading required package: carData
##
## Attaching package: 'car'
##
## The following object is masked from 'package:arm':
##
## logit
##
## The following object is masked from 'package:dplyr':
##
## recode
##
## The following object is masked from 'package:purrr':
##
## some
library(survey)
## Loading required package: grid
## Loading required package: survival
##
## Attaching package: 'survey'
##
## The following object is masked from 'package:sjstats':
##
## cv
##
## The following object is masked from 'package:Hmisc':
##
## deff
##
## The following object is masked from 'package:graphics':
##
## dotchart
library(forcats)
## 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 = 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")
mint11_distinct<-mint11_df %>% distinct(caseid)
## NEW REGRESSION analysis wrangling
glimpse(mint11_df_1$p_party_new)
## dbl+lbl [1:13816] 3, 1, 4, 4, 1, 1, 1, 4, 1, 1, 1, 2, 3, 1, 1, 3, 3, 2, 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() %>%filter(!is.na(hub_dist)) %>%
mutate(p_party_new = case_when(p_party == 1 ~ 1,
p_party == 2 ~ 2,
T ~ 3),
p_party_new = labelled(p_party_new, 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)),
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 = 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)),
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))
# 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 = as.factor(p_party_new)) %>%
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))
glimpse(mint11_df_1$religion_in_new)
## Factor w/ 3 levels "1","2","3": 2 1 3 3 1 1 1 3 1 1 ...
mint11_df_1 <- mint11_df_1 %>%
mutate(p_party_new = relevel(p_party_new, ref = 3),#change in reference level
religion_in_new= relevel(religion_in_new, ref= 3))
## q9_1 DV <- age + partisanship + city tier + region (2) + religion + party.
model_q9_1 <-mint11_df_1 %>% glm(q9 ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+geo_region_in_cb_new+religion_in_new, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_q9_1$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","SE","Hindu","Muslim")
summary(model_q9_1)
##
## Call:
## glm(formula = q9 ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + geo_region_in_cb_new + religion_in_new, family = binomial,
## data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept -0.48064 0.10309 -4.662 3.13e-06 ***
## Partisan_Index -0.11518 0.02543 -4.529 5.92e-06 ***
## age 0.22916 0.02693 8.511 < 2e-16 ***
## Rest_Tiers -0.05155 0.04001 -1.289 0.1975
## BJP -0.59208 0.04855 -12.195 < 2e-16 ***
## INC 0.01836 0.06707 0.274 0.7842
## SE -0.00148 0.03983 -0.037 0.9704
## Hindu -0.10103 0.05116 -1.975 0.0483 *
## Muslim 1.44835 0.07602 19.053 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 16605 on 12433 degrees of freedom
## Residual deviance: 15441 on 12425 degrees of freedom
## AIC: 15280
##
## Number of Fisher Scoring iterations: 4
summ(model = model_q9_1, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
12434
|
|
Dependent variable
|
q9
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(8)
|
1164.14
|
|
Pseudo-R² (Cragg-Uhler)
|
0.13
|
|
Pseudo-R² (McFadden)
|
0.07
|
|
AIC
|
15280.34
|
|
BIC
|
15347.20
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.62
|
0.51
|
0.76
|
-4.66
|
0.00
|
|
Partisan_Index
|
0.89
|
0.85
|
0.94
|
-4.53
|
0.00
|
|
age
|
1.26
|
1.19
|
1.33
|
8.51
|
0.00
|
|
Rest_Tiers
|
0.95
|
0.88
|
1.03
|
-1.29
|
0.20
|
|
BJP
|
0.55
|
0.50
|
0.61
|
-12.20
|
0.00
|
|
INC
|
1.02
|
0.89
|
1.16
|
0.27
|
0.78
|
|
SE
|
1.00
|
0.92
|
1.08
|
-0.04
|
0.97
|
|
Hindu
|
0.90
|
0.82
|
1.00
|
-1.97
|
0.05
|
|
Muslim
|
4.26
|
3.67
|
4.94
|
19.05
|
0.00
|
|
Standard errors: MLE
|
vif(model_q9_1)
## Warning in vif.default(model_q9_1): No intercept: vifs may not be sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 7.356719 1 2.712327
## age_cb_new 11.225335 1 3.350423
## geo_citytier_in 2.653795 1 1.629047
## p_party_new 3.551306 2 1.372768
## geo_region_in_cb_new 1.919723 1 1.385541
## religion_in_new 5.800143 2 1.551886
## q9_2 DV <- age + partisanship + city tier + region (4) + religion + party
model_q9_2 <-mint11_df_1 %>% glm(q9 ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+geo_region_in_cb+religion_in_new, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_q9_2$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","South India", "East & Northeast India", "West India","Hindu","Muslim")
summary(model_q9_2)
##
## Call:
## glm(formula = q9 ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + geo_region_in_cb + religion_in_new, family = binomial,
## data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept -0.510087 0.105958 -4.814 1.48e-06 ***
## Partisan_Index -0.117944 0.025468 -4.631 3.64e-06 ***
## age 0.229635 0.026983 8.511 < 2e-16 ***
## Rest_Tiers -0.053113 0.040067 -1.326 0.185
## BJP -0.581292 0.048719 -11.932 < 2e-16 ***
## INC 0.009682 0.067189 0.144 0.885
## South India 0.082099 0.050215 1.635 0.102
## East & Northeast India -0.094430 0.058463 -1.615 0.106
## West India 0.044267 0.054518 0.812 0.417
## Hindu -0.084089 0.051498 -1.633 0.102
## Muslim 1.457529 0.076129 19.145 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 16605 on 12433 degrees of freedom
## Residual deviance: 15432 on 12423 degrees of freedom
## AIC: 15277
##
## Number of Fisher Scoring iterations: 4
summ(model = model_q9_2, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
12434
|
|
Dependent variable
|
q9
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(10)
|
1173.06
|
|
Pseudo-R² (Cragg-Uhler)
|
0.13
|
|
Pseudo-R² (McFadden)
|
0.07
|
|
AIC
|
15276.68
|
|
BIC
|
15358.39
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.60
|
0.49
|
0.74
|
-4.81
|
0.00
|
|
Partisan_Index
|
0.89
|
0.85
|
0.93
|
-4.63
|
0.00
|
|
age
|
1.26
|
1.19
|
1.33
|
8.51
|
0.00
|
|
Rest_Tiers
|
0.95
|
0.88
|
1.03
|
-1.33
|
0.18
|
|
BJP
|
0.56
|
0.51
|
0.62
|
-11.93
|
0.00
|
|
INC
|
1.01
|
0.89
|
1.15
|
0.14
|
0.89
|
|
South India
|
1.09
|
0.98
|
1.20
|
1.63
|
0.10
|
|
East & Northeast India
|
0.91
|
0.81
|
1.02
|
-1.62
|
0.11
|
|
West India
|
1.05
|
0.94
|
1.16
|
0.81
|
0.42
|
|
Hindu
|
0.92
|
0.83
|
1.02
|
-1.63
|
0.10
|
|
Muslim
|
4.30
|
3.70
|
4.99
|
19.15
|
0.00
|
|
Standard errors: MLE
|
vif(model_q9_2)
## Warning in vif.default(model_q9_2): No intercept: vifs may not be sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 7.371162 1 2.714988
## age_cb_new 11.263001 1 3.356040
## geo_citytier_in 2.658315 1 1.630434
## p_party_new 3.600242 2 1.377472
## geo_region_in_cb 3.135560 3 1.209816
## religion_in_new 5.876445 2 1.556964
## q9_3 DV <- age + partisanship + city tier + distance + religion + party.
model_q9_3 <-mint11_df_1 %>% glm(q9 ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+hub_dist+religion_in_new, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_q9_3$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","Distance","Hindu","Muslim")
summary(model_q9_3)
##
## Call:
## glm(formula = q9 ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + hub_dist + religion_in_new, family = binomial,
## data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept -4.555e-01 1.076e-01 -4.233 2.30e-05 ***
## Partisan_Index -1.153e-01 2.543e-02 -4.535 5.76e-06 ***
## age 2.289e-01 2.693e-02 8.501 < 2e-16 ***
## Rest_Tiers -4.853e-02 4.020e-02 -1.207 0.2274
## BJP -5.976e-01 4.855e-02 -12.309 < 2e-16 ***
## INC 1.880e-02 6.707e-02 0.280 0.7792
## Distance -2.407e-05 3.572e-05 -0.674 0.5004
## Hindu -1.025e-01 5.119e-02 -2.003 0.0452 *
## Muslim 1.447e+00 7.601e-02 19.035 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 16605 on 12433 degrees of freedom
## Residual deviance: 15440 on 12425 degrees of freedom
## AIC: 15280
##
## Number of Fisher Scoring iterations: 4
summ(model = model_q9_3, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
12434
|
|
Dependent variable
|
q9
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(8)
|
1164.59
|
|
Pseudo-R² (Cragg-Uhler)
|
0.13
|
|
Pseudo-R² (McFadden)
|
0.07
|
|
AIC
|
15280.43
|
|
BIC
|
15347.28
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.63
|
0.51
|
0.78
|
-4.23
|
0.00
|
|
Partisan_Index
|
0.89
|
0.85
|
0.94
|
-4.54
|
0.00
|
|
age
|
1.26
|
1.19
|
1.33
|
8.50
|
0.00
|
|
Rest_Tiers
|
0.95
|
0.88
|
1.03
|
-1.21
|
0.23
|
|
BJP
|
0.55
|
0.50
|
0.61
|
-12.31
|
0.00
|
|
INC
|
1.02
|
0.89
|
1.16
|
0.28
|
0.78
|
|
Distance
|
1.00
|
1.00
|
1.00
|
-0.67
|
0.50
|
|
Hindu
|
0.90
|
0.82
|
1.00
|
-2.00
|
0.05
|
|
Muslim
|
4.25
|
3.66
|
4.93
|
19.04
|
0.00
|
|
Standard errors: MLE
|
vif(model_q9_3)
## Warning in vif.default(model_q9_3): No intercept: vifs may not be sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 7.354956 1 2.712002
## age_cb_new 11.226722 1 3.350630
## geo_citytier_in 2.680762 1 1.637303
## p_party_new 3.554333 2 1.373060
## hub_dist 4.254037 1 2.062532
## religion_in_new 5.804994 2 1.552210
## q9_4 DV <- age + partisanship + city tier + bordering + religion + party
model_q9_4 <-mint11_df_1 %>% glm(q9 ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+bordering+religion_in_new, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_q9_4$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","bordering","Hindu","Muslim")
summary(model_q9_4)
##
## Call:
## glm(formula = q9 ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + bordering + religion_in_new, family = binomial,
## data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept -0.27919 0.17712 -1.576 0.1150
## Partisan_Index -0.11534 0.02543 -4.536 5.73e-06 ***
## age 0.22910 0.02693 8.507 < 2e-16 ***
## Rest_Tiers -0.06511 0.04113 -1.583 0.1134
## BJP -0.59635 0.04788 -12.454 < 2e-16 ***
## INC 0.01671 0.06708 0.249 0.8032
## bordering -0.09959 0.07180 -1.387 0.1654
## Hindu -0.10140 0.05114 -1.983 0.0474 *
## Muslim 1.44795 0.07598 19.056 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 16605 on 12433 degrees of freedom
## Residual deviance: 15439 on 12425 degrees of freedom
## AIC: 15279
##
## Number of Fisher Scoring iterations: 4
summ(model = model_q9_4, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
12434
|
|
Dependent variable
|
q9
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(8)
|
1166.05
|
|
Pseudo-R² (Cragg-Uhler)
|
0.13
|
|
Pseudo-R² (McFadden)
|
0.07
|
|
AIC
|
15278.90
|
|
BIC
|
15345.75
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.76
|
0.53
|
1.07
|
-1.58
|
0.11
|
|
Partisan_Index
|
0.89
|
0.85
|
0.94
|
-4.54
|
0.00
|
|
age
|
1.26
|
1.19
|
1.33
|
8.51
|
0.00
|
|
Rest_Tiers
|
0.94
|
0.86
|
1.02
|
-1.58
|
0.11
|
|
BJP
|
0.55
|
0.50
|
0.61
|
-12.45
|
0.00
|
|
INC
|
1.02
|
0.89
|
1.16
|
0.25
|
0.80
|
|
bordering
|
0.91
|
0.79
|
1.04
|
-1.39
|
0.17
|
|
Hindu
|
0.90
|
0.82
|
1.00
|
-1.98
|
0.05
|
|
Muslim
|
4.25
|
3.67
|
4.94
|
19.06
|
0.00
|
|
Standard errors: MLE
|
vif(model_q9_4)
## Warning in vif.default(model_q9_4): No intercept: vifs may not be sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 7.354025 1 2.711831
## age_cb_new 11.227881 1 3.350803
## geo_citytier_in 2.804248 1 1.674589
## p_party_new 3.437414 2 1.361626
## bordering 51.923818 1 7.205818
## religion_in_new 5.792726 2 1.551389
# q10 Important
## q10 wrangling
mint11_df_1<-mint11_df_1 %>% mutate(q10=ifelse(q10 ==2, 0, q10))
## q10_1 DV <- age + partisanship + city tier + region (2) + religion + party.
model_q10_1 <-mint11_df_1 %>% glm(q10 ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+geo_region_in_cb_new+religion_in_new, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_q10_1$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","SE","Hindu","Muslim")
summary(model_q10_1)
##
## Call:
## glm(formula = q10 ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + geo_region_in_cb_new + religion_in_new, family = binomial,
## data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept -0.17354 0.10386 -1.671 0.09474 .
## Partisan_Index -0.47869 0.02589 -18.488 < 2e-16 ***
## age 0.17814 0.02693 6.615 3.73e-11 ***
## Rest_Tiers -0.11174 0.03997 -2.795 0.00518 **
## BJP 0.19445 0.04835 4.022 5.78e-05 ***
## INC -0.07058 0.07019 -1.006 0.31460
## SE -0.23225 0.04004 -5.800 6.63e-09 ***
## Hindu 0.34646 0.05463 6.342 2.26e-10 ***
## Muslim -0.39098 0.08003 -4.885 1.03e-06 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 16283 on 12433 degrees of freedom
## Residual deviance: 15355 on 12425 degrees of freedom
## AIC: 15260
##
## Number of Fisher Scoring iterations: 4
summ(model = model_q10_1, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
12434
|
|
Dependent variable
|
q10
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(8)
|
928.33
|
|
Pseudo-R² (Cragg-Uhler)
|
0.10
|
|
Pseudo-R² (McFadden)
|
0.06
|
|
AIC
|
15259.63
|
|
BIC
|
15326.48
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.84
|
0.69
|
1.03
|
-1.67
|
0.09
|
|
Partisan_Index
|
0.62
|
0.59
|
0.65
|
-18.49
|
0.00
|
|
age
|
1.19
|
1.13
|
1.26
|
6.61
|
0.00
|
|
Rest_Tiers
|
0.89
|
0.83
|
0.97
|
-2.80
|
0.01
|
|
BJP
|
1.21
|
1.10
|
1.34
|
4.02
|
0.00
|
|
INC
|
0.93
|
0.81
|
1.07
|
-1.01
|
0.31
|
|
SE
|
0.79
|
0.73
|
0.86
|
-5.80
|
0.00
|
|
Hindu
|
1.41
|
1.27
|
1.57
|
6.34
|
0.00
|
|
Muslim
|
0.68
|
0.58
|
0.79
|
-4.89
|
0.00
|
|
Standard errors: MLE
|
vif(model_q10_1)
## Warning in vif.default(model_q10_1): No intercept: vifs may not be sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 6.708071 1 2.589994
## age_cb_new 11.074052 1 3.327770
## geo_citytier_in 2.612881 1 1.616441
## p_party_new 4.079882 2 1.421222
## geo_region_in_cb_new 1.818259 1 1.348428
## religion_in_new 7.179438 2 1.636902
## q10_2 DV <- age + partisanship + city tier + region (4) + religion + party
model_q10_2 <-mint11_df_1 %>% glm(q10 ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+geo_region_in_cb+religion_in_new, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_q10_2$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","South India", "East & Northeast India", "West India","Hindu","Muslim")
summary(model_q10_2)
##
## Call:
## glm(formula = q10 ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + geo_region_in_cb + religion_in_new, family = binomial,
## data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept -0.18049 0.10684 -1.689 0.091132 .
## Partisan_Index -0.47779 0.02599 -18.385 < 2e-16 ***
## age 0.18281 0.02702 6.767 1.32e-11 ***
## Rest_Tiers -0.10444 0.04008 -2.606 0.009163 **
## BJP 0.16571 0.04861 3.409 0.000652 ***
## INC -0.04374 0.07055 -0.620 0.535297
## South India -0.38592 0.05192 -7.433 1.06e-13 ***
## East & Northeast India 0.05704 0.05653 1.009 0.312944
## West India 0.09258 0.05343 1.733 0.083137 .
## Hindu 0.31421 0.05504 5.709 1.14e-08 ***
## Muslim -0.41731 0.08042 -5.189 2.11e-07 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 16283 on 12433 degrees of freedom
## Residual deviance: 15300 on 12423 degrees of freedom
## AIC: 15222
##
## Number of Fisher Scoring iterations: 4
summ(model = model_q10_2, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
12434
|
|
Dependent variable
|
q10
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(10)
|
982.76
|
|
Pseudo-R² (Cragg-Uhler)
|
0.10
|
|
Pseudo-R² (McFadden)
|
0.06
|
|
AIC
|
15222.03
|
|
BIC
|
15303.74
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.83
|
0.68
|
1.03
|
-1.69
|
0.09
|
|
Partisan_Index
|
0.62
|
0.59
|
0.65
|
-18.38
|
0.00
|
|
age
|
1.20
|
1.14
|
1.27
|
6.77
|
0.00
|
|
Rest_Tiers
|
0.90
|
0.83
|
0.97
|
-2.61
|
0.01
|
|
BJP
|
1.18
|
1.07
|
1.30
|
3.41
|
0.00
|
|
INC
|
0.96
|
0.83
|
1.10
|
-0.62
|
0.54
|
|
South India
|
0.68
|
0.61
|
0.75
|
-7.43
|
0.00
|
|
East & Northeast India
|
1.06
|
0.95
|
1.18
|
1.01
|
0.31
|
|
West India
|
1.10
|
0.99
|
1.22
|
1.73
|
0.08
|
|
Hindu
|
1.37
|
1.23
|
1.53
|
5.71
|
0.00
|
|
Muslim
|
0.66
|
0.56
|
0.77
|
-5.19
|
0.00
|
|
Standard errors: MLE
|
vif(model_q10_2)
## Warning in vif.default(model_q10_2): No intercept: vifs may not be sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 6.728810 1 2.593995
## age_cb_new 11.090504 1 3.330241
## geo_citytier_in 2.610539 1 1.615716
## p_party_new 4.143417 2 1.426723
## geo_region_in_cb 2.917674 3 1.195380
## religion_in_new 7.271457 2 1.642122
## q10_3 DV <- age + partisanship + city tier + distance + religion + party.
model_q10_3 <-mint11_df_1 %>% glm(q10 ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+hub_dist+religion_in_new, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_q10_3$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","Distance","Hindu","Muslim")
summary(model_q10_3)
##
## Call:
## glm(formula = q10 ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + hub_dist + religion_in_new, family = binomial,
## data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept -7.679e-02 1.085e-01 -0.708 0.4790
## Partisan_Index -4.774e-01 2.588e-02 -18.442 < 2e-16 ***
## age 1.774e-01 2.693e-02 6.587 4.50e-11 ***
## Rest_Tiers -9.883e-02 4.012e-02 -2.463 0.0138 *
## BJP 1.936e-01 4.834e-02 4.006 6.19e-05 ***
## INC -6.838e-02 7.020e-02 -0.974 0.3300
## Distance -2.140e-04 3.601e-05 -5.944 2.79e-09 ***
## Hindu 3.412e-01 5.466e-02 6.243 4.29e-10 ***
## Muslim -3.914e-01 8.004e-02 -4.889 1.01e-06 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 16283 on 12433 degrees of freedom
## Residual deviance: 15353 on 12425 degrees of freedom
## AIC: 15256
##
## Number of Fisher Scoring iterations: 4
summ(model = model_q10_3, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
12434
|
|
Dependent variable
|
q10
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(8)
|
930.01
|
|
Pseudo-R² (Cragg-Uhler)
|
0.10
|
|
Pseudo-R² (McFadden)
|
0.06
|
|
AIC
|
15256.28
|
|
BIC
|
15323.13
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.93
|
0.75
|
1.15
|
-0.71
|
0.48
|
|
Partisan_Index
|
0.62
|
0.59
|
0.65
|
-18.44
|
0.00
|
|
age
|
1.19
|
1.13
|
1.26
|
6.59
|
0.00
|
|
Rest_Tiers
|
0.91
|
0.84
|
0.98
|
-2.46
|
0.01
|
|
BJP
|
1.21
|
1.10
|
1.33
|
4.01
|
0.00
|
|
INC
|
0.93
|
0.81
|
1.07
|
-0.97
|
0.33
|
|
Distance
|
1.00
|
1.00
|
1.00
|
-5.94
|
0.00
|
|
Hindu
|
1.41
|
1.26
|
1.57
|
6.24
|
0.00
|
|
Muslim
|
0.68
|
0.58
|
0.79
|
-4.89
|
0.00
|
|
Standard errors: MLE
|
vif(model_q10_3)
## Warning in vif.default(model_q10_3): No intercept: vifs may not be sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 6.703338 1 2.589080
## age_cb_new 11.072770 1 3.327577
## geo_citytier_in 2.631271 1 1.622119
## p_party_new 4.080832 2 1.421305
## hub_dist 4.072195 1 2.017968
## religion_in_new 7.186250 2 1.637290
## q10_4 DV <- age + partisanship + city tier + bordering + religion + party
model_q10_4 <-mint11_df_1 %>% glm(q10 ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+bordering+religion_in_new, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_q10_4$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","bordering","Hindu","Muslim")
summary(model_q10_4)
##
## Call:
## glm(formula = q10 ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + bordering + religion_in_new, family = binomial,
## data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept -0.18551 0.17720 -1.047 0.2951
## Partisan_Index -0.47535 0.02583 -18.400 < 2e-16 ***
## age 0.17997 0.02691 6.689 2.25e-11 ***
## Rest_Tiers -0.13131 0.04110 -3.195 0.0014 **
## BJP 0.24243 0.04759 5.094 3.51e-07 ***
## INC -0.07109 0.07004 -1.015 0.3101
## bordering -0.06102 0.07152 -0.853 0.3936
## Hindu 0.35274 0.05456 6.465 1.01e-10 ***
## Muslim -0.37922 0.07990 -4.746 2.07e-06 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 16283 on 12433 degrees of freedom
## Residual deviance: 15388 on 12425 degrees of freedom
## AIC: 15296
##
## Number of Fisher Scoring iterations: 4
summ(model = model_q10_4, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
12434
|
|
Dependent variable
|
q10
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(8)
|
895.35
|
|
Pseudo-R² (Cragg-Uhler)
|
0.09
|
|
Pseudo-R² (McFadden)
|
0.05
|
|
AIC
|
15295.81
|
|
BIC
|
15362.66
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.83
|
0.59
|
1.18
|
-1.05
|
0.30
|
|
Partisan_Index
|
0.62
|
0.59
|
0.65
|
-18.40
|
0.00
|
|
age
|
1.20
|
1.14
|
1.26
|
6.69
|
0.00
|
|
Rest_Tiers
|
0.88
|
0.81
|
0.95
|
-3.20
|
0.00
|
|
BJP
|
1.27
|
1.16
|
1.40
|
5.09
|
0.00
|
|
INC
|
0.93
|
0.81
|
1.07
|
-1.02
|
0.31
|
|
bordering
|
0.94
|
0.82
|
1.08
|
-0.85
|
0.39
|
|
Hindu
|
1.42
|
1.28
|
1.58
|
6.46
|
0.00
|
|
Muslim
|
0.68
|
0.59
|
0.80
|
-4.75
|
0.00
|
|
Standard errors: MLE
|
vif(model_q10_4)
## Warning in vif.default(model_q10_4): No intercept: vifs may not be sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 6.692747 1 2.587034
## age_cb_new 11.087933 1 3.329855
## geo_citytier_in 2.773342 1 1.665335
## p_party_new 3.951039 2 1.409866
## bordering 51.381389 1 7.168081
## religion_in_new 7.182202 2 1.637059
# q10 Partisan and party interaction
## q10_5 to q10_8 wrangling
mint11_df_1<- mint11_df_1 %>% mutate(p_party_new_1=case_when(p_party_new==1~1,
TRUE~0),
partisan_index_new = (q_p_party_1_new + q_p_party_2_new +
q_p_party_3_new + q_p_party_4_new)/4,
partisan_index_party=partisan_index_new*p_party_new_1)
## q10_5 DV <- age + city tier + region (4) + religion + partisanship*party.
model_q10_5 <-mint11_df_1 %>% glm(q10 ~ partisan_index+age_cb_new+ geo_citytier_in+partisan_index_party+p_party_new+geo_region_in_cb_new+religion_in_new, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_q10_5$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","partisan*party","BJP","INC","SE","Hindu","Muslim")
summary(model_q10_5)
##
## Call:
## glm(formula = q10 ~ partisan_index + age_cb_new + geo_citytier_in +
## partisan_index_party + p_party_new + geo_region_in_cb_new +
## religion_in_new, family = binomial, data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept -0.31068 0.11263 -2.758 0.00581 **
## Partisan_Index -0.42232 0.03134 -13.478 < 2e-16 ***
## age 0.18072 0.02696 6.702 2.06e-11 ***
## Rest_Tiers -0.11066 0.04000 -2.767 0.00566 **
## partisan*party -0.13402 0.04215 -3.179 0.00148 **
## BJP 0.50364 0.10854 4.640 3.48e-06 ***
## INC -0.03512 0.07090 -0.495 0.62032
## SE -0.22884 0.04007 -5.711 1.13e-08 ***
## Hindu 0.35116 0.05461 6.430 1.27e-10 ***
## Muslim -0.38041 0.07996 -4.757 1.96e-06 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 16283 on 12433 degrees of freedom
## Residual deviance: 15344 on 12424 degrees of freedom
## AIC: 15252
##
## Number of Fisher Scoring iterations: 4
summ(model = model_q10_5, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
12434
|
|
Dependent variable
|
q10
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(9)
|
938.49
|
|
Pseudo-R² (Cragg-Uhler)
|
0.10
|
|
Pseudo-R² (McFadden)
|
0.06
|
|
AIC
|
15251.71
|
|
BIC
|
15325.99
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.73
|
0.59
|
0.91
|
-2.76
|
0.01
|
|
Partisan_Index
|
0.66
|
0.62
|
0.70
|
-13.48
|
0.00
|
|
age
|
1.20
|
1.14
|
1.26
|
6.70
|
0.00
|
|
Rest_Tiers
|
0.90
|
0.83
|
0.97
|
-2.77
|
0.01
|
|
partisan*party
|
0.87
|
0.81
|
0.95
|
-3.18
|
0.00
|
|
BJP
|
1.65
|
1.34
|
2.05
|
4.64
|
0.00
|
|
INC
|
0.97
|
0.84
|
1.11
|
-0.50
|
0.62
|
|
SE
|
0.80
|
0.74
|
0.86
|
-5.71
|
0.00
|
|
Hindu
|
1.42
|
1.28
|
1.58
|
6.43
|
0.00
|
|
Muslim
|
0.68
|
0.58
|
0.80
|
-4.76
|
0.00
|
|
Standard errors: MLE
|
vif(model_q10_5)
## Warning in vif.default(model_q10_5): No intercept: vifs may not be sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 9.866159 1 3.141044
## age_cb_new 11.091759 1 3.330429
## geo_citytier_in 2.612131 1 1.616209
## partisan_index_party 11.249422 1 3.354016
## p_party_new 21.729333 2 2.159045
## geo_region_in_cb_new 1.821451 1 1.349611
## religion_in_new 7.154792 2 1.635495
## q10_6 DV <- age + city tier + region (2) + religion + partisanship*party
model_q10_6 <-mint11_df_1 %>% glm(q10 ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+partisan_index_party+geo_region_in_cb+religion_in_new, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_q10_6$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","Partisan index*party","South India", "East & Northeast India", "West India","Hindu","Muslim")
summary(model_q10_6)
##
## Call:
## glm(formula = q10 ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + partisan_index_party + geo_region_in_cb + religion_in_new,
## family = binomial, data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept -0.321491 0.115599 -2.781 0.00542 **
## Partisan_Index -0.420408 0.031451 -13.367 < 2e-16 ***
## age 0.185551 0.027050 6.860 6.90e-12 ***
## Rest_Tiers -0.103151 0.040104 -2.572 0.01011 *
## BJP 0.480348 0.108925 4.410 1.03e-05 ***
## INC -0.007299 0.071280 -0.102 0.91845
## Partisan index*party -0.136302 0.042263 -3.225 0.00126 **
## South India -0.381758 0.051942 -7.350 1.99e-13 ***
## East & Northeast India 0.062092 0.056610 1.097 0.27271
## West India 0.094521 0.053492 1.767 0.07723 .
## Hindu 0.318976 0.055024 5.797 6.75e-09 ***
## Muslim -0.406381 0.080340 -5.058 4.23e-07 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 16283 on 12433 degrees of freedom
## Residual deviance: 15290 on 12422 degrees of freedom
## AIC: 15214
##
## Number of Fisher Scoring iterations: 4
summ(model = model_q10_6, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
12434
|
|
Dependent variable
|
q10
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(11)
|
993.22
|
|
Pseudo-R² (Cragg-Uhler)
|
0.10
|
|
Pseudo-R² (McFadden)
|
0.06
|
|
AIC
|
15213.82
|
|
BIC
|
15302.96
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.73
|
0.58
|
0.91
|
-2.78
|
0.01
|
|
Partisan_Index
|
0.66
|
0.62
|
0.70
|
-13.37
|
0.00
|
|
age
|
1.20
|
1.14
|
1.27
|
6.86
|
0.00
|
|
Rest_Tiers
|
0.90
|
0.83
|
0.98
|
-2.57
|
0.01
|
|
BJP
|
1.62
|
1.31
|
2.00
|
4.41
|
0.00
|
|
INC
|
0.99
|
0.86
|
1.14
|
-0.10
|
0.92
|
|
Partisan index*party
|
0.87
|
0.80
|
0.95
|
-3.23
|
0.00
|
|
South India
|
0.68
|
0.62
|
0.76
|
-7.35
|
0.00
|
|
East & Northeast India
|
1.06
|
0.95
|
1.19
|
1.10
|
0.27
|
|
West India
|
1.10
|
0.99
|
1.22
|
1.77
|
0.08
|
|
Hindu
|
1.38
|
1.24
|
1.53
|
5.80
|
0.00
|
|
Muslim
|
0.67
|
0.57
|
0.78
|
-5.06
|
0.00
|
|
Standard errors: MLE
|
vif(model_q10_6)
## Warning in vif.default(model_q10_6): No intercept: vifs may not be sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 9.895268 1 3.145675
## age_cb_new 11.108764 1 3.332981
## geo_citytier_in 2.609546 1 1.615409
## p_party_new 21.927792 2 2.163957
## partisan_index_party 11.272607 1 3.357470
## geo_region_in_cb 2.924706 3 1.195860
## religion_in_new 7.246752 2 1.640725
## q10_7 DV <- age + city tier + bordering + religion + partisanship*party.
model_q10_4 <-mint11_df_1 %>% glm(q10 ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+partisan_index_party+bordering+religion_in_new, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_q10_4$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","Partisanship*party", "bordering","Hindu","Muslim")
summary(model_q10_4)
##
## Call:
## glm(formula = q10 ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + partisan_index_party + bordering + religion_in_new,
## family = binomial, data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept -0.33492 0.18313 -1.829 0.067422 .
## Partisan_Index -0.41659 0.03126 -13.326 < 2e-16 ***
## age 0.18261 0.02694 6.779 1.21e-11 ***
## Rest_Tiers -0.12957 0.04113 -3.151 0.001629 **
## BJP 0.56438 0.10791 5.230 1.69e-07 ***
## INC -0.03429 0.07073 -0.485 0.627836
## Partisanship*party -0.13983 0.04209 -3.322 0.000893 ***
## bordering -0.05681 0.07166 -0.793 0.427956
## Hindu 0.35762 0.05454 6.557 5.51e-11 ***
## Muslim -0.36827 0.07982 -4.614 3.95e-06 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 16283 on 12433 degrees of freedom
## Residual deviance: 15376 on 12424 degrees of freedom
## AIC: 15287
##
## Number of Fisher Scoring iterations: 4
summ(model = model_q10_4, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
12434
|
|
Dependent variable
|
q10
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(9)
|
906.45
|
|
Pseudo-R² (Cragg-Uhler)
|
0.09
|
|
Pseudo-R² (McFadden)
|
0.05
|
|
AIC
|
15286.91
|
|
BIC
|
15361.20
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.72
|
0.50
|
1.02
|
-1.83
|
0.07
|
|
Partisan_Index
|
0.66
|
0.62
|
0.70
|
-13.33
|
0.00
|
|
age
|
1.20
|
1.14
|
1.27
|
6.78
|
0.00
|
|
Rest_Tiers
|
0.88
|
0.81
|
0.95
|
-3.15
|
0.00
|
|
BJP
|
1.76
|
1.42
|
2.17
|
5.23
|
0.00
|
|
INC
|
0.97
|
0.84
|
1.11
|
-0.48
|
0.63
|
|
Partisanship*party
|
0.87
|
0.80
|
0.94
|
-3.32
|
0.00
|
|
bordering
|
0.94
|
0.82
|
1.09
|
-0.79
|
0.43
|
|
Hindu
|
1.43
|
1.28
|
1.59
|
6.56
|
0.00
|
|
Muslim
|
0.69
|
0.59
|
0.81
|
-4.61
|
0.00
|
|
Standard errors: MLE
|
vif(model_q10_4)
## Warning in vif.default(model_q10_4): No intercept: vifs may not be sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 9.843300 1 3.137403
## age_cb_new 11.105399 1 3.332476
## geo_citytier_in 2.772418 1 1.665058
## p_party_new 21.531632 2 2.154117
## partisan_index_party 11.247815 1 3.353776
## bordering 51.512449 1 7.177217
## religion_in_new 7.155096 2 1.635512
## q10_8 DV <- age + city tier + distance + religion + partisanship*party.
model_q10_4 <-mint11_df_1 %>% glm(q10 ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+partisan_index_party+hub_dist+religion_in_new, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_q10_4$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","Partisanship*party", "distance","Hindu","Muslim")
summary(model_q10_4)
##
## Call:
## glm(formula = q10 ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + partisan_index_party + hub_dist + religion_in_new,
## family = binomial, data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept -2.155e-01 1.170e-01 -1.841 0.06558 .
## Partisan_Index -4.209e-01 3.133e-02 -13.434 < 2e-16 ***
## age 1.800e-01 2.696e-02 6.675 2.47e-11 ***
## Rest_Tiers -9.786e-02 4.015e-02 -2.437 0.01479 *
## BJP 5.035e-01 1.085e-01 4.640 3.48e-06 ***
## INC -3.286e-02 7.091e-02 -0.463 0.64310
## Partisanship*party -1.344e-01 4.215e-02 -3.188 0.00143 **
## distance -2.112e-04 3.603e-05 -5.861 4.60e-09 ***
## Hindu 3.460e-01 5.464e-02 6.332 2.41e-10 ***
## Muslim -3.808e-01 7.997e-02 -4.761 1.92e-06 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 16283 on 12433 degrees of freedom
## Residual deviance: 15343 on 12424 degrees of freedom
## AIC: 15248
##
## Number of Fisher Scoring iterations: 4
summ(model = model_q10_4, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
12434
|
|
Dependent variable
|
q10
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(9)
|
940.22
|
|
Pseudo-R² (Cragg-Uhler)
|
0.10
|
|
Pseudo-R² (McFadden)
|
0.06
|
|
AIC
|
15248.30
|
|
BIC
|
15322.58
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.81
|
0.64
|
1.01
|
-1.84
|
0.07
|
|
Partisan_Index
|
0.66
|
0.62
|
0.70
|
-13.43
|
0.00
|
|
age
|
1.20
|
1.14
|
1.26
|
6.67
|
0.00
|
|
Rest_Tiers
|
0.91
|
0.84
|
0.98
|
-2.44
|
0.01
|
|
BJP
|
1.65
|
1.34
|
2.05
|
4.64
|
0.00
|
|
INC
|
0.97
|
0.84
|
1.11
|
-0.46
|
0.64
|
|
Partisanship*party
|
0.87
|
0.80
|
0.95
|
-3.19
|
0.00
|
|
distance
|
1.00
|
1.00
|
1.00
|
-5.86
|
0.00
|
|
Hindu
|
1.41
|
1.27
|
1.57
|
6.33
|
0.00
|
|
Muslim
|
0.68
|
0.58
|
0.80
|
-4.76
|
0.00
|
|
Standard errors: MLE
|
vif(model_q10_4)
## Warning in vif.default(model_q10_4): No intercept: vifs may not be sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 9.860005 1 3.140064
## age_cb_new 11.090594 1 3.330254
## geo_citytier_in 2.630725 1 1.621951
## p_party_new 21.726393 2 2.158971
## partisan_index_party 11.248985 1 3.353951
## hub_dist 4.077974 1 2.019399
## religion_in_new 7.161398 2 1.635873
# q11a
## wrangling
mint11_df_1<-mint11_df_1 %>% mutate(q11a=ifelse(q11a ==2, 0, q11a))
## q11a_1 DV <- age + partisanship + city tier + region (2) + religion + party.
model_q11a_1 <-mint11_df_1 %>% glm(q11a ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+geo_region_in_cb_new+religion_in_new, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_q11a_1$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","SE","Hindu","Muslim")
summary(model_q11a_1)
##
## Call:
## glm(formula = q11a ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + geo_region_in_cb_new + religion_in_new, family = binomial,
## data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept 1.73190 0.17557 9.864 < 2e-16 ***
## Partisan_Index -0.26644 0.04261 -6.253 4.04e-10 ***
## age -0.21433 0.04511 -4.752 2.02e-06 ***
## Rest_Tiers -0.10652 0.06459 -1.649 0.09912 .
## BJP 0.11965 0.07847 1.525 0.12733
## INC 0.10857 0.12017 0.903 0.36626
## SE -0.15107 0.06481 -2.331 0.01975 *
## Hindu -0.10633 0.09348 -1.138 0.25532
## Muslim -0.41584 0.14098 -2.950 0.00318 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5967.2 on 4563 degrees of freedom
## Residual deviance: 5861.1 on 4555 degrees of freedom
## (7870 observations deleted due to missingness)
## AIC: 5718.7
##
## Number of Fisher Scoring iterations: 4
summ(model = model_q11a_1, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
4564 (7870 missing obs. deleted)
|
|
Dependent variable
|
q11a
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(8)
|
106.06
|
|
Pseudo-R² (Cragg-Uhler)
|
0.03
|
|
Pseudo-R² (McFadden)
|
0.02
|
|
AIC
|
5718.70
|
|
BIC
|
5776.54
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
5.65
|
4.01
|
7.97
|
9.86
|
0.00
|
|
Partisan_Index
|
0.77
|
0.70
|
0.83
|
-6.25
|
0.00
|
|
age
|
0.81
|
0.74
|
0.88
|
-4.75
|
0.00
|
|
Rest_Tiers
|
0.90
|
0.79
|
1.02
|
-1.65
|
0.10
|
|
BJP
|
1.13
|
0.97
|
1.31
|
1.52
|
0.13
|
|
INC
|
1.11
|
0.88
|
1.41
|
0.90
|
0.37
|
|
SE
|
0.86
|
0.76
|
0.98
|
-2.33
|
0.02
|
|
Hindu
|
0.90
|
0.75
|
1.08
|
-1.14
|
0.26
|
|
Muslim
|
0.66
|
0.50
|
0.87
|
-2.95
|
0.00
|
|
Standard errors: MLE
|
vif(model_q11a_1)
## Warning in vif.default(model_q11a_1): No intercept: vifs may not be sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 6.381596 1 2.526182
## age_cb_new 12.406934 1 3.522348
## geo_citytier_in 2.612986 1 1.616473
## p_party_new 4.366042 2 1.445513
## geo_region_in_cb_new 1.731430 1 1.315838
## religion_in_new 8.352838 2 1.700038
## q11a_2 DV <- age + partisanship + city tier + region (4) + religion + party
model_q11a_2 <-mint11_df_1 %>% glm(q11a ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+geo_region_in_cb+religion_in_new, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_q11a_2$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","South India", "East & Northeast India", "West India","Hindu","Muslim")
summary(model_q11a_2)
##
## Call:
## glm(formula = q11a ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + geo_region_in_cb + religion_in_new, family = binomial,
## data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept 1.79390 0.18041 9.944 < 2e-16 ***
## Partisan_Index -0.26490 0.04272 -6.201 5.61e-10 ***
## age -0.21751 0.04524 -4.807 1.53e-06 ***
## Rest_Tiers -0.11308 0.06470 -1.748 0.08051 .
## BJP 0.10834 0.07889 1.373 0.16966
## INC 0.11960 0.12048 0.993 0.32083
## South India -0.26971 0.08640 -3.122 0.00180 **
## East & Northeast India -0.08920 0.08899 -1.002 0.31617
## West India -0.08818 0.08487 -1.039 0.29885
## Hindu -0.12391 0.09394 -1.319 0.18718
## Muslim -0.42860 0.14128 -3.034 0.00242 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5967.2 on 4563 degrees of freedom
## Residual deviance: 5856.8 on 4553 degrees of freedom
## (7870 observations deleted due to missingness)
## AIC: 5717.4
##
## Number of Fisher Scoring iterations: 4
summ(model = model_q11a_2, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
4564 (7870 missing obs. deleted)
|
|
Dependent variable
|
q11a
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(10)
|
110.35
|
|
Pseudo-R² (Cragg-Uhler)
|
0.03
|
|
Pseudo-R² (McFadden)
|
0.02
|
|
AIC
|
5717.42
|
|
BIC
|
5788.11
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
6.01
|
4.22
|
8.56
|
9.94
|
0.00
|
|
Partisan_Index
|
0.77
|
0.71
|
0.83
|
-6.20
|
0.00
|
|
age
|
0.80
|
0.74
|
0.88
|
-4.81
|
0.00
|
|
Rest_Tiers
|
0.89
|
0.79
|
1.01
|
-1.75
|
0.08
|
|
BJP
|
1.11
|
0.95
|
1.30
|
1.37
|
0.17
|
|
INC
|
1.13
|
0.89
|
1.43
|
0.99
|
0.32
|
|
South India
|
0.76
|
0.64
|
0.90
|
-3.12
|
0.00
|
|
East & Northeast India
|
0.91
|
0.77
|
1.09
|
-1.00
|
0.32
|
|
West India
|
0.92
|
0.78
|
1.08
|
-1.04
|
0.30
|
|
Hindu
|
0.88
|
0.73
|
1.06
|
-1.32
|
0.19
|
|
Muslim
|
0.65
|
0.49
|
0.86
|
-3.03
|
0.00
|
|
Standard errors: MLE
|
vif(model_q11a_2)
## Warning in vif.default(model_q11a_2): No intercept: vifs may not be sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 6.405152 1 2.530840
## age_cb_new 12.467360 1 3.530915
## geo_citytier_in 2.618527 1 1.618186
## p_party_new 4.452907 2 1.452650
## geo_region_in_cb 2.842938 3 1.190222
## religion_in_new 8.438911 2 1.704400
## q11a_3 DV <- age + partisanship + city tier + distance + religion + party.
model_q11a_3 <-mint11_df_1 %>% glm(q11a ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+hub_dist+religion_in_new, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_q11a_3$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","Distance","Hindu","Muslim")
summary(model_q11a_3)
##
## Call:
## glm(formula = q11a ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + hub_dist + religion_in_new, family = binomial,
## data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept 1.750e+00 1.831e-01 9.561 < 2e-16 ***
## Partisan_Index -2.639e-01 4.258e-02 -6.197 5.74e-10 ***
## age -2.146e-01 4.510e-02 -4.758 1.95e-06 ***
## Rest_Tiers -1.013e-01 6.465e-02 -1.566 0.11729
## BJP 1.279e-01 7.842e-02 1.631 0.10286
## INC 1.088e-01 1.201e-01 0.906 0.36519
## Distance -9.791e-05 5.947e-05 -1.646 0.09968 .
## Hindu -1.093e-01 9.353e-02 -1.168 0.24277
## Muslim -4.151e-01 1.409e-01 -2.945 0.00323 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5967.2 on 4563 degrees of freedom
## Residual deviance: 5863.8 on 4555 degrees of freedom
## (7870 observations deleted due to missingness)
## AIC: 5722.2
##
## Number of Fisher Scoring iterations: 4
summ(model = model_q10_3, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
12434
|
|
Dependent variable
|
q10
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(8)
|
930.01
|
|
Pseudo-R² (Cragg-Uhler)
|
0.10
|
|
Pseudo-R² (McFadden)
|
0.06
|
|
AIC
|
15256.28
|
|
BIC
|
15323.13
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.93
|
0.75
|
1.15
|
-0.71
|
0.48
|
|
Partisan_Index
|
0.62
|
0.59
|
0.65
|
-18.44
|
0.00
|
|
age
|
1.19
|
1.13
|
1.26
|
6.59
|
0.00
|
|
Rest_Tiers
|
0.91
|
0.84
|
0.98
|
-2.46
|
0.01
|
|
BJP
|
1.21
|
1.10
|
1.33
|
4.01
|
0.00
|
|
INC
|
0.93
|
0.81
|
1.07
|
-0.97
|
0.33
|
|
Distance
|
1.00
|
1.00
|
1.00
|
-5.94
|
0.00
|
|
Hindu
|
1.41
|
1.26
|
1.57
|
6.24
|
0.00
|
|
Muslim
|
0.68
|
0.58
|
0.79
|
-4.89
|
0.00
|
|
Standard errors: MLE
|
vif(model_q11a_3)
## Warning in vif.default(model_q11a_3): No intercept: vifs may not be sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 6.375037 1 2.524884
## age_cb_new 12.409201 1 3.522670
## geo_citytier_in 2.619481 1 1.618481
## p_party_new 4.365085 2 1.445434
## hub_dist 4.001093 1 2.000273
## religion_in_new 8.367648 2 1.700791
## q11a_4 DV <- age + partisanship + city tier + bordering + religion + party
model_q11_4 <-mint11_df_1 %>% glm(q11a ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+bordering+religion_in_new, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_q11_4$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","bordering","Hindu","Muslim")
summary(model_q11_4)
##
## Call:
## glm(formula = q11a ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + bordering + religion_in_new, family = binomial,
## data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept 1.45488 0.28299 5.141 2.73e-07 ***
## Partisan_Index -0.26434 0.04259 -6.206 5.42e-10 ***
## age -0.21256 0.04507 -4.716 2.41e-06 ***
## Rest_Tiers -0.09281 0.06664 -1.393 0.1638
## BJP 0.14848 0.07759 1.914 0.0557 .
## INC 0.10754 0.12009 0.895 0.3705
## bordering 0.09493 0.11239 0.845 0.3983
## Hindu -0.10239 0.09337 -1.097 0.2728
## Muslim -0.41402 0.14091 -2.938 0.0033 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5967.2 on 4563 degrees of freedom
## Residual deviance: 5865.8 on 4555 degrees of freedom
## (7870 observations deleted due to missingness)
## AIC: 5726.8
##
## Number of Fisher Scoring iterations: 4
summ(model = model_q11_4, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
4564 (7870 missing obs. deleted)
|
|
Dependent variable
|
q11a
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(8)
|
101.35
|
|
Pseudo-R² (Cragg-Uhler)
|
0.03
|
|
Pseudo-R² (McFadden)
|
0.02
|
|
AIC
|
5726.80
|
|
BIC
|
5784.63
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
4.28
|
2.46
|
7.46
|
5.14
|
0.00
|
|
Partisan_Index
|
0.77
|
0.71
|
0.83
|
-6.21
|
0.00
|
|
age
|
0.81
|
0.74
|
0.88
|
-4.72
|
0.00
|
|
Rest_Tiers
|
0.91
|
0.80
|
1.04
|
-1.39
|
0.16
|
|
BJP
|
1.16
|
1.00
|
1.35
|
1.91
|
0.06
|
|
INC
|
1.11
|
0.88
|
1.41
|
0.90
|
0.37
|
|
bordering
|
1.10
|
0.88
|
1.37
|
0.84
|
0.40
|
|
Hindu
|
0.90
|
0.75
|
1.08
|
-1.10
|
0.27
|
|
Muslim
|
0.66
|
0.50
|
0.87
|
-2.94
|
0.00
|
|
Standard errors: MLE
|
vif(model_q11_4)
## Warning in vif.default(model_q11_4): No intercept: vifs may not be sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 6.381086 1 2.526081
## age_cb_new 12.400774 1 3.521473
## geo_citytier_in 2.785399 1 1.668952
## p_party_new 4.251763 2 1.435960
## bordering 48.185946 1 6.941610
## religion_in_new 8.345633 2 1.699671
## q11a_5 DV <- age + city tier + region (4) + religion + partisanship*party.
model_q11a_5 <-mint11_df_1 %>% glm(q11a ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+partisan_index_party+geo_region_in_cb+religion_in_new, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_q11a_5$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","Partisan index*party","South India", "East & Northeast India", "West India","Hindu","Muslim")
summary(model_q11a_5)
##
## Call:
## glm(formula = q11a ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + partisan_index_party + geo_region_in_cb + religion_in_new,
## family = binomial, data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept 1.86602 0.19470 9.584 < 2e-16 ***
## Partisan_Index -0.29514 0.05243 -5.629 1.81e-08 ***
## age -0.21905 0.04528 -4.838 1.31e-06 ***
## Rest_Tiers -0.11431 0.06472 -1.766 0.07738 .
## BJP -0.04170 0.16948 -0.246 0.80566
## INC 0.09969 0.12227 0.815 0.41488
## Partisan index*party 0.06678 0.06675 1.001 0.31705
## South India -0.27211 0.08646 -3.147 0.00165 **
## East & Northeast India -0.09168 0.08903 -1.030 0.30315
## West India -0.08912 0.08488 -1.050 0.29372
## Hindu -0.12547 0.09402 -1.335 0.18202
## Muslim -0.43332 0.14148 -3.063 0.00219 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5967.2 on 4563 degrees of freedom
## Residual deviance: 5855.8 on 4552 degrees of freedom
## (7870 observations deleted due to missingness)
## AIC: 5719.3
##
## Number of Fisher Scoring iterations: 4
summ(model = model_q11a_5, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
4564 (7870 missing obs. deleted)
|
|
Dependent variable
|
q11a
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(11)
|
111.35
|
|
Pseudo-R² (Cragg-Uhler)
|
0.03
|
|
Pseudo-R² (McFadden)
|
0.02
|
|
AIC
|
5719.29
|
|
BIC
|
5796.40
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
6.46
|
4.41
|
9.47
|
9.58
|
0.00
|
|
Partisan_Index
|
0.74
|
0.67
|
0.82
|
-5.63
|
0.00
|
|
age
|
0.80
|
0.74
|
0.88
|
-4.84
|
0.00
|
|
Rest_Tiers
|
0.89
|
0.79
|
1.01
|
-1.77
|
0.08
|
|
BJP
|
0.96
|
0.69
|
1.34
|
-0.25
|
0.81
|
|
INC
|
1.10
|
0.87
|
1.40
|
0.82
|
0.41
|
|
Partisan index*party
|
1.07
|
0.94
|
1.22
|
1.00
|
0.32
|
|
South India
|
0.76
|
0.64
|
0.90
|
-3.15
|
0.00
|
|
East & Northeast India
|
0.91
|
0.77
|
1.09
|
-1.03
|
0.30
|
|
West India
|
0.91
|
0.77
|
1.08
|
-1.05
|
0.29
|
|
Hindu
|
0.88
|
0.73
|
1.06
|
-1.33
|
0.18
|
|
Muslim
|
0.65
|
0.49
|
0.86
|
-3.06
|
0.00
|
|
Standard errors: MLE
|
vif(model_q11a_5)
## Warning in vif.default(model_q11a_5): No intercept: vifs may not be sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 9.635607 1 3.104127
## age_cb_new 12.483126 1 3.533147
## geo_citytier_in 2.619639 1 1.618530
## p_party_new 21.896019 2 2.163173
## partisan_index_party 10.970861 1 3.312229
## geo_region_in_cb 2.845694 3 1.190414
## religion_in_new 8.460173 2 1.705473
## q11a_6 DV <- age + city tier + region (2) + religion + partisanship*party
model_q11a_6 <-mint11_df_1 %>% glm(q11a ~ partisan_index+age_cb_new+ geo_citytier_in+partisan_index_party+p_party_new+geo_region_in_cb_new+religion_in_new, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_q11a_6$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","partisan*party","BJP","INC","SE","Hindu","Muslim")
summary(model_q11a_6)
##
## Call:
## glm(formula = q11a ~ partisan_index + age_cb_new + geo_citytier_in +
## partisan_index_party + p_party_new + geo_region_in_cb_new +
## religion_in_new, family = binomial, data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept 1.80274 0.18996 9.490 < 2e-16 ***
## Partisan_Index -0.29640 0.05234 -5.663 1.49e-08 ***
## age -0.21578 0.04514 -4.780 1.75e-06 ***
## Rest_Tiers -0.10767 0.06461 -1.667 0.09560 .
## partisan*party 0.06617 0.06673 0.991 0.32145
## BJP -0.02902 0.16928 -0.171 0.86387
## INC 0.08884 0.12196 0.728 0.46632
## SE -0.15311 0.06485 -2.361 0.01823 *
## Hindu -0.10785 0.09355 -1.153 0.24898
## Muslim -0.42056 0.14119 -2.979 0.00289 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5967.2 on 4563 degrees of freedom
## Residual deviance: 5860.1 on 4554 degrees of freedom
## (7870 observations deleted due to missingness)
## AIC: 5720.6
##
## Number of Fisher Scoring iterations: 4
summ(model = model_q11a_6, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
4564 (7870 missing obs. deleted)
|
|
Dependent variable
|
q11a
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(9)
|
107.04
|
|
Pseudo-R² (Cragg-Uhler)
|
0.03
|
|
Pseudo-R² (McFadden)
|
0.02
|
|
AIC
|
5720.55
|
|
BIC
|
5784.81
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
6.07
|
4.18
|
8.80
|
9.49
|
0.00
|
|
Partisan_Index
|
0.74
|
0.67
|
0.82
|
-5.66
|
0.00
|
|
age
|
0.81
|
0.74
|
0.88
|
-4.78
|
0.00
|
|
Rest_Tiers
|
0.90
|
0.79
|
1.02
|
-1.67
|
0.10
|
|
partisan*party
|
1.07
|
0.94
|
1.22
|
0.99
|
0.32
|
|
BJP
|
0.97
|
0.70
|
1.35
|
-0.17
|
0.86
|
|
INC
|
1.09
|
0.86
|
1.39
|
0.73
|
0.47
|
|
SE
|
0.86
|
0.76
|
0.97
|
-2.36
|
0.02
|
|
Hindu
|
0.90
|
0.75
|
1.08
|
-1.15
|
0.25
|
|
Muslim
|
0.66
|
0.50
|
0.87
|
-2.98
|
0.00
|
|
Standard errors: MLE
|
vif(model_q11a_6)
## Warning in vif.default(model_q11a_6): No intercept: vifs may not be sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 9.613851 1 3.100621
## age_cb_new 12.419863 1 3.524183
## geo_citytier_in 2.613998 1 1.616786
## partisan_index_party 10.974319 1 3.312751
## p_party_new 21.740268 2 2.159316
## geo_region_in_cb_new 1.732761 1 1.316344
## religion_in_new 8.374698 2 1.701149
## q11a_7 DV <- age + city tier + bordering + religion + partisanship*party.
model_q11a_7 <-mint11_df_1 %>% glm(q11a ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+partisan_index_party+bordering+religion_in_new, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_q11a_7$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","Partisanship*party", "bordering","Hindu","Muslim")
summary(model_q11a_7)
##
## Call:
## glm(formula = q11a ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + partisan_index_party + bordering + religion_in_new,
## family = binomial, data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept 1.52195 0.29283 5.197 2.02e-07 ***
## Partisan_Index -0.29151 0.05224 -5.580 2.41e-08 ***
## age -0.21386 0.04510 -4.742 2.12e-06 ***
## Rest_Tiers -0.09406 0.06667 -1.411 0.15826
## BJP 0.01353 0.16847 0.080 0.93599
## INC 0.08953 0.12187 0.735 0.46257
## Partisanship*party 0.06020 0.06669 0.903 0.36671
## bordering 0.09304 0.11241 0.828 0.40787
## Hindu -0.10374 0.09344 -1.110 0.26688
## Muslim -0.41821 0.14110 -2.964 0.00304 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5967.2 on 4563 degrees of freedom
## Residual deviance: 5865.0 on 4554 degrees of freedom
## (7870 observations deleted due to missingness)
## AIC: 5728.7
##
## Number of Fisher Scoring iterations: 4
summ(model = model_q11a_7, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
4564 (7870 missing obs. deleted)
|
|
Dependent variable
|
q11a
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(9)
|
102.16
|
|
Pseudo-R² (Cragg-Uhler)
|
0.03
|
|
Pseudo-R² (McFadden)
|
0.02
|
|
AIC
|
5728.75
|
|
BIC
|
5793.01
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
4.58
|
2.58
|
8.13
|
5.20
|
0.00
|
|
Partisan_Index
|
0.75
|
0.67
|
0.83
|
-5.58
|
0.00
|
|
age
|
0.81
|
0.74
|
0.88
|
-4.74
|
0.00
|
|
Rest_Tiers
|
0.91
|
0.80
|
1.04
|
-1.41
|
0.16
|
|
BJP
|
1.01
|
0.73
|
1.41
|
0.08
|
0.94
|
|
INC
|
1.09
|
0.86
|
1.39
|
0.73
|
0.46
|
|
Partisanship*party
|
1.06
|
0.93
|
1.21
|
0.90
|
0.37
|
|
bordering
|
1.10
|
0.88
|
1.37
|
0.83
|
0.41
|
|
Hindu
|
0.90
|
0.75
|
1.08
|
-1.11
|
0.27
|
|
Muslim
|
0.66
|
0.50
|
0.87
|
-2.96
|
0.00
|
|
Standard errors: MLE
|
vif(model_q11a_7)
## Warning in vif.default(model_q11a_7): No intercept: vifs may not be sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 9.589556 1 3.096701
## age_cb_new 12.413453 1 3.523273
## geo_citytier_in 2.786692 1 1.669339
## p_party_new 21.522747 2 2.153894
## partisan_index_party 10.970968 1 3.312245
## bordering 48.195265 1 6.942281
## religion_in_new 8.365985 2 1.700706
## q11a_8 DV <- age + city tier + distance + religion + partisanship*party.
model_q11a_8 <-mint11_df_1 %>% glm(q11a ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+partisan_index_party+hub_dist+religion_in_new, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_q11a_8$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","Partisanship*party", "distance","Hindu","Muslim")
summary(model_q11a_8)
##
## Call:
## glm(formula = q11a ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + partisan_index_party + hub_dist + religion_in_new,
## family = binomial, data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept 1.819e+00 1.970e-01 9.233 < 2e-16 ***
## Partisan_Index -2.928e-01 5.227e-02 -5.601 2.13e-08 ***
## age -2.160e-01 4.513e-02 -4.786 1.70e-06 ***
## Rest_Tiers -1.023e-01 6.466e-02 -1.582 0.11370
## BJP -1.560e-02 1.691e-01 -0.092 0.92650
## INC 8.974e-02 1.219e-01 0.736 0.46168
## Partisanship*party 6.391e-02 6.669e-02 0.958 0.33790
## distance -9.931e-05 5.949e-05 -1.669 0.09508 .
## Hindu -1.108e-01 9.361e-02 -1.184 0.23658
## Muslim -4.197e-01 1.411e-01 -2.973 0.00295 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 5967.2 on 4563 degrees of freedom
## Residual deviance: 5862.9 on 4554 degrees of freedom
## (7870 observations deleted due to missingness)
## AIC: 5724.1
##
## Number of Fisher Scoring iterations: 4
summ(model = model_q11a_8, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
4564 (7870 missing obs. deleted)
|
|
Dependent variable
|
q11a
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(9)
|
104.27
|
|
Pseudo-R² (Cragg-Uhler)
|
0.03
|
|
Pseudo-R² (McFadden)
|
0.02
|
|
AIC
|
5724.10
|
|
BIC
|
5788.36
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
6.17
|
4.19
|
9.07
|
9.23
|
0.00
|
|
Partisan_Index
|
0.75
|
0.67
|
0.83
|
-5.60
|
0.00
|
|
age
|
0.81
|
0.74
|
0.88
|
-4.79
|
0.00
|
|
Rest_Tiers
|
0.90
|
0.80
|
1.02
|
-1.58
|
0.11
|
|
BJP
|
0.98
|
0.71
|
1.37
|
-0.09
|
0.93
|
|
INC
|
1.09
|
0.86
|
1.39
|
0.74
|
0.46
|
|
Partisanship*party
|
1.07
|
0.94
|
1.21
|
0.96
|
0.34
|
|
distance
|
1.00
|
1.00
|
1.00
|
-1.67
|
0.10
|
|
Hindu
|
0.90
|
0.75
|
1.08
|
-1.18
|
0.24
|
|
Muslim
|
0.66
|
0.50
|
0.87
|
-2.97
|
0.00
|
|
Standard errors: MLE
|
vif(model_q11a_8)
## Warning in vif.default(model_q11a_8): No intercept: vifs may not be sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 9.596459 1 3.097815
## age_cb_new 12.422175 1 3.524511
## geo_citytier_in 2.620185 1 1.618699
## p_party_new 21.706314 2 2.158473
## partisan_index_party 10.967963 1 3.311791
## hub_dist 4.002316 1 2.000579
## religion_in_new 8.389398 2 1.701895
# q11b narrow morality
## data wrangling
mint11_df_4<-mint11_df_1 %>%filter(!is.na(q11b)) %>% mutate(q11b_narrow=case_when(q11b==2~1,
TRUE~0))
model_q11b_narrow_2 <-mint11_df_4 %>% glm(q11b_narrow ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+geo_region_in_cb+religion_in_new, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_q11b_narrow_2$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","South India", "East & Northeast India", "West India","Hindu","Muslim")
summary(model_q11b_narrow_2)
##
## Call:
## glm(formula = q11b_narrow ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + geo_region_in_cb + religion_in_new, family = binomial,
## data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept -1.00142 0.13980 -7.163 7.87e-13 ***
## Partisan_Index 0.27758 0.03405 8.153 3.54e-16 ***
## age -0.10130 0.03495 -2.899 0.003746 **
## Rest_Tiers -0.08825 0.05366 -1.645 0.100023
## BJP -0.19365 0.06642 -2.916 0.003549 **
## INC 0.08736 0.08367 1.044 0.296467
## South India 0.11438 0.06474 1.767 0.077284 .
## East & Northeast India -0.18202 0.08311 -2.190 0.028511 *
## West India -0.03456 0.07533 -0.459 0.646394
## Hindu -0.19018 0.06531 -2.912 0.003591 **
## Muslim -0.29323 0.08675 -3.380 0.000725 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 9135.2 on 7869 degrees of freedom
## Residual deviance: 8943.9 on 7859 degrees of freedom
## AIC: 9072.8
##
## Number of Fisher Scoring iterations: 4
summ(model = model_q11b_narrow_2, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
7870
|
|
Dependent variable
|
q11b_narrow
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(10)
|
191.34
|
|
Pseudo-R² (Cragg-Uhler)
|
0.03
|
|
Pseudo-R² (McFadden)
|
0.02
|
|
AIC
|
9072.76
|
|
BIC
|
9149.44
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.37
|
0.28
|
0.48
|
-7.16
|
0.00
|
|
Partisan_Index
|
1.32
|
1.23
|
1.41
|
8.15
|
0.00
|
|
age
|
0.90
|
0.84
|
0.97
|
-2.90
|
0.00
|
|
Rest_Tiers
|
0.92
|
0.82
|
1.02
|
-1.64
|
0.10
|
|
BJP
|
0.82
|
0.72
|
0.94
|
-2.92
|
0.00
|
|
INC
|
1.09
|
0.93
|
1.29
|
1.04
|
0.30
|
|
South India
|
1.12
|
0.99
|
1.27
|
1.77
|
0.08
|
|
East & Northeast India
|
0.83
|
0.71
|
0.98
|
-2.19
|
0.03
|
|
West India
|
0.97
|
0.83
|
1.12
|
-0.46
|
0.65
|
|
Hindu
|
0.83
|
0.73
|
0.94
|
-2.91
|
0.00
|
|
Muslim
|
0.75
|
0.63
|
0.88
|
-3.38
|
0.00
|
|
Standard errors: MLE
|
vif(model_q11b_narrow_2)
## Warning in vif.default(model_q11b_narrow_2): No intercept: vifs may not be
## sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 9.083606 1 3.013902
## age_cb_new 10.163201 1 3.187978
## geo_citytier_in 2.697222 1 1.642322
## p_party_new 3.144600 2 1.331654
## geo_region_in_cb 3.432651 3 1.228207
## religion_in_new 5.060923 2 1.499883
## q11b_2 narrow morality
model_q11b_2 <-mint11_df_4 %>% glm(q11b_narrow ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+partisan_index_party+geo_region_in_cb+religion_in_new, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_q11b_2$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","Partisan index*party","South India", "East & Northeast India", "West India","Hindu","Muslim")
summary(model_q11b_2)
##
## Call:
## glm(formula = q11b_narrow ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + partisan_index_party + geo_region_in_cb + religion_in_new,
## family = binomial, data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept -1.02560 0.15185 -6.754 1.44e-11 ***
## Partisan_Index 0.28645 0.04038 7.094 1.30e-12 ***
## age -0.10091 0.03496 -2.886 0.003899 **
## Rest_Tiers -0.08816 0.05366 -1.643 0.100376
## BJP -0.13026 0.16811 -0.775 0.438428
## INC 0.09313 0.08489 1.097 0.272605
## Partisan index*party -0.02510 0.06117 -0.410 0.681544
## South India 0.11524 0.06478 1.779 0.075259 .
## East & Northeast India -0.18115 0.08313 -2.179 0.029332 *
## West India -0.03421 0.07533 -0.454 0.649715
## Hindu -0.18959 0.06534 -2.902 0.003710 **
## Muslim -0.29170 0.08685 -3.359 0.000783 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 9135.2 on 7869 degrees of freedom
## Residual deviance: 8943.7 on 7858 degrees of freedom
## AIC: 9074.7
##
## Number of Fisher Scoring iterations: 4
summ(model = model_q11b_2, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
7870
|
|
Dependent variable
|
q11b_narrow
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(11)
|
191.50
|
|
Pseudo-R² (Cragg-Uhler)
|
0.03
|
|
Pseudo-R² (McFadden)
|
0.02
|
|
AIC
|
9074.72
|
|
BIC
|
9158.36
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.36
|
0.27
|
0.48
|
-6.75
|
0.00
|
|
Partisan_Index
|
1.33
|
1.23
|
1.44
|
7.09
|
0.00
|
|
age
|
0.90
|
0.84
|
0.97
|
-2.89
|
0.00
|
|
Rest_Tiers
|
0.92
|
0.82
|
1.02
|
-1.64
|
0.10
|
|
BJP
|
0.88
|
0.63
|
1.22
|
-0.77
|
0.44
|
|
INC
|
1.10
|
0.93
|
1.30
|
1.10
|
0.27
|
|
Partisan index*party
|
0.98
|
0.87
|
1.10
|
-0.41
|
0.68
|
|
South India
|
1.12
|
0.99
|
1.27
|
1.78
|
0.08
|
|
East & Northeast India
|
0.83
|
0.71
|
0.98
|
-2.18
|
0.03
|
|
West India
|
0.97
|
0.83
|
1.12
|
-0.45
|
0.65
|
|
Hindu
|
0.83
|
0.73
|
0.94
|
-2.90
|
0.00
|
|
Muslim
|
0.75
|
0.63
|
0.89
|
-3.36
|
0.00
|
|
Standard errors: MLE
|
vif(model_q11b_2)
## Warning in vif.default(model_q11b_2): No intercept: vifs may not be sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 12.773882 1 3.574057
## age_cb_new 10.170736 1 3.189159
## geo_citytier_in 2.697054 1 1.642271
## p_party_new 21.110668 2 2.143510
## partisan_index_party 11.674641 1 3.416817
## geo_region_in_cb 3.436435 3 1.228433
## religion_in_new 5.072183 2 1.500717
# 11b wider morality
## data wrangling
mint11_df_4<-mint11_df_1 %>%filter(!is.na(q11b)) %>% mutate(q11b_wider=case_when(q11b%in%c(2,3,6)~1,
TRUE~0))
model_q11b_wider_1 <-mint11_df_4 %>% glm(q11b_wider ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+geo_region_in_cb+religion_in_new, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_q11b_wider_1$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","South India", "East & Northeast India", "West India","Hindu","Muslim")
summary(model_q11b_wider_1)
##
## Call:
## glm(formula = q11b_wider ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + geo_region_in_cb + religion_in_new, family = binomial,
## data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept -0.05906 0.12623 -0.468 0.63989
## Partisan_Index 0.09211 0.03036 3.033 0.00242 **
## age -0.01601 0.03162 -0.506 0.61275
## Rest_Tiers -0.09466 0.04847 -1.953 0.05082 .
## BJP -0.26833 0.05913 -4.538 5.67e-06 ***
## INC -0.13501 0.07678 -1.758 0.07868 .
## South India 0.19373 0.05899 3.284 0.00102 **
## East & Northeast India -0.06325 0.07229 -0.875 0.38158
## West India 0.01822 0.06729 0.271 0.78655
## Hindu -0.19849 0.06085 -3.262 0.00111 **
## Muslim 0.16137 0.07931 2.035 0.04190 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 10593 on 7869 degrees of freedom
## Residual deviance: 10437 on 7859 degrees of freedom
## AIC: 10522
##
## Number of Fisher Scoring iterations: 4
summ(model = model_q11b_wider_1, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
7870
|
|
Dependent variable
|
q11b_wider
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(10)
|
156.32
|
|
Pseudo-R² (Cragg-Uhler)
|
0.03
|
|
Pseudo-R² (McFadden)
|
0.01
|
|
AIC
|
10521.73
|
|
BIC
|
10598.41
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.94
|
0.74
|
1.21
|
-0.47
|
0.64
|
|
Partisan_Index
|
1.10
|
1.03
|
1.16
|
3.03
|
0.00
|
|
age
|
0.98
|
0.92
|
1.05
|
-0.51
|
0.61
|
|
Rest_Tiers
|
0.91
|
0.83
|
1.00
|
-1.95
|
0.05
|
|
BJP
|
0.76
|
0.68
|
0.86
|
-4.54
|
0.00
|
|
INC
|
0.87
|
0.75
|
1.02
|
-1.76
|
0.08
|
|
South India
|
1.21
|
1.08
|
1.36
|
3.28
|
0.00
|
|
East & Northeast India
|
0.94
|
0.81
|
1.08
|
-0.87
|
0.38
|
|
West India
|
1.02
|
0.89
|
1.16
|
0.27
|
0.79
|
|
Hindu
|
0.82
|
0.73
|
0.92
|
-3.26
|
0.00
|
|
Muslim
|
1.18
|
1.01
|
1.37
|
2.03
|
0.04
|
|
Standard errors: MLE
|
vif(model_q11b_wider_1)
## Warning in vif.default(model_q11b_wider_1): No intercept: vifs may not be
## sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 8.228167 1 2.868478
## age_cb_new 10.386986 1 3.222885
## geo_citytier_in 2.744338 1 1.656604
## p_party_new 3.379798 2 1.355884
## geo_region_in_cb 3.363120 3 1.224025
## religion_in_new 5.704229 2 1.545430
## q11b_2 wider morality
model_q11b_wider_2 <-mint11_df_4 %>% glm(q11b_wider ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+partisan_index_party+geo_region_in_cb+religion_in_new, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_q11b_wider_2$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","Partisan index*party","South India", "East & Northeast India", "West India","Hindu","Muslim")
summary(model_q11b_wider_2)
##
## Call:
## glm(formula = q11b_wider ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + partisan_index_party + geo_region_in_cb + religion_in_new,
## family = binomial, data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept -0.12312 0.13652 -0.902 0.367133
## Partisan_Index 0.11605 0.03604 3.220 0.001282 **
## age -0.01482 0.03164 -0.468 0.639529
## Rest_Tiers -0.09449 0.04848 -1.949 0.051283 .
## BJP -0.10556 0.14433 -0.731 0.464529
## INC -0.11968 0.07781 -1.538 0.124026
## Partisan index*party -0.06613 0.05352 -1.236 0.216591
## South India 0.19615 0.05903 3.323 0.000891 ***
## East & Northeast India -0.06074 0.07232 -0.840 0.400940
## West India 0.01937 0.06731 0.288 0.773462
## Hindu -0.19660 0.06088 -3.229 0.001241 **
## Muslim 0.16598 0.07943 2.090 0.036651 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 10593 on 7869 degrees of freedom
## Residual deviance: 10435 on 7858 degrees of freedom
## AIC: 10523
##
## Number of Fisher Scoring iterations: 4
summ(model = model_q11b_wider_2, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
7870
|
|
Dependent variable
|
q11b_wider
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(11)
|
157.85
|
|
Pseudo-R² (Cragg-Uhler)
|
0.03
|
|
Pseudo-R² (McFadden)
|
0.01
|
|
AIC
|
10522.91
|
|
BIC
|
10606.56
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.88
|
0.68
|
1.16
|
-0.90
|
0.37
|
|
Partisan_Index
|
1.12
|
1.05
|
1.21
|
3.22
|
0.00
|
|
age
|
0.99
|
0.93
|
1.05
|
-0.47
|
0.64
|
|
Rest_Tiers
|
0.91
|
0.83
|
1.00
|
-1.95
|
0.05
|
|
BJP
|
0.90
|
0.68
|
1.19
|
-0.73
|
0.46
|
|
INC
|
0.89
|
0.76
|
1.03
|
-1.54
|
0.12
|
|
Partisan index*party
|
0.94
|
0.84
|
1.04
|
-1.24
|
0.22
|
|
South India
|
1.22
|
1.08
|
1.37
|
3.32
|
0.00
|
|
East & Northeast India
|
0.94
|
0.82
|
1.08
|
-0.84
|
0.40
|
|
West India
|
1.02
|
0.89
|
1.16
|
0.29
|
0.77
|
|
Hindu
|
0.82
|
0.73
|
0.93
|
-3.23
|
0.00
|
|
Muslim
|
1.18
|
1.01
|
1.38
|
2.09
|
0.04
|
|
Standard errors: MLE
|
vif(model_q11b_wider_2)
## Warning in vif.default(model_q11b_wider_2): No intercept: vifs may not be
## sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 11.573288 1 3.401954
## age_cb_new 10.395252 1 3.224167
## geo_citytier_in 2.744730 1 1.656723
## p_party_new 21.132180 2 2.144056
## partisan_index_party 11.437146 1 3.381885
## geo_region_in_cb 3.367140 3 1.224269
## religion_in_new 5.721244 2 1.546581
# Set 5: (Q10 < dissatisfied with terrorism)
## DV <- age + partisanship + city tier + region (4) + religion + party + dissatisfied.
model_q10_2 <-mint11_df_1 %>% glm(q10 ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+geo_region_in_cb+religion_in_new+q3_4, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_q10_2$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","South India", "East & Northeast India", "West India","Hindu","Muslim","Terrorism_Dissatisfied")
summary(model_q10_2)
##
## Call:
## glm(formula = q10 ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + geo_region_in_cb + religion_in_new + q3_4,
## family = binomial, data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept 0.05046 0.11532 0.438 0.66166
## Partisan_Index -0.46237 0.02616 -17.674 < 2e-16 ***
## age 0.17783 0.02707 6.570 5.03e-11 ***
## Rest_Tiers -0.10682 0.04013 -2.662 0.00777 **
## BJP 0.10527 0.04991 2.109 0.03494 *
## INC -0.01252 0.07094 -0.176 0.85993
## South India -0.38718 0.05199 -7.447 9.54e-14 ***
## East & Northeast India 0.06054 0.05660 1.070 0.28482
## West India 0.09307 0.05348 1.740 0.08180 .
## Hindu 0.28333 0.05539 5.115 3.14e-07 ***
## Muslim -0.37653 0.08096 -4.651 3.30e-06 ***
## Terrorism_Dissatisfied -0.10237 0.01909 -5.363 8.20e-08 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 16283 on 12433 degrees of freedom
## Residual deviance: 15271 on 12422 degrees of freedom
## AIC: 15192
##
## Number of Fisher Scoring iterations: 4
summ(model = model_q10_2, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
12434
|
|
Dependent variable
|
q10
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(11)
|
1011.82
|
|
Pseudo-R² (Cragg-Uhler)
|
0.10
|
|
Pseudo-R² (McFadden)
|
0.06
|
|
AIC
|
15191.54
|
|
BIC
|
15280.68
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
1.05
|
0.84
|
1.32
|
0.44
|
0.66
|
|
Partisan_Index
|
0.63
|
0.60
|
0.66
|
-17.67
|
0.00
|
|
age
|
1.19
|
1.13
|
1.26
|
6.57
|
0.00
|
|
Rest_Tiers
|
0.90
|
0.83
|
0.97
|
-2.66
|
0.01
|
|
BJP
|
1.11
|
1.01
|
1.23
|
2.11
|
0.03
|
|
INC
|
0.99
|
0.86
|
1.13
|
-0.18
|
0.86
|
|
South India
|
0.68
|
0.61
|
0.75
|
-7.45
|
0.00
|
|
East & Northeast India
|
1.06
|
0.95
|
1.19
|
1.07
|
0.28
|
|
West India
|
1.10
|
0.99
|
1.22
|
1.74
|
0.08
|
|
Hindu
|
1.33
|
1.19
|
1.48
|
5.12
|
0.00
|
|
Muslim
|
0.69
|
0.59
|
0.80
|
-4.65
|
0.00
|
|
Terrorism_Dissatisfied
|
0.90
|
0.87
|
0.94
|
-5.36
|
0.00
|
|
Standard errors: MLE
|
vif(model_q10_2)
## Warning in vif.default(model_q10_2): No intercept: vifs may not be sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 6.805665 1 2.608767
## age_cb_new 11.109713 1 3.333124
## geo_citytier_in 2.611327 1 1.615960
## p_party_new 4.484380 2 1.455210
## geo_region_in_cb 2.916772 3 1.195319
## religion_in_new 7.531587 2 1.656615
## q3_4 4.793630 1 2.189436
## DV <- age + partisanship + city tier + region (2) + religion + party + dissatisfied.
model_q10_1 <-mint11_df_1 %>% glm(q10 ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+geo_region_in_cb_new+religion_in_new+q3_4, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_q10_1$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","SE","Hindu","Muslim","Terrorism_Dissatisfied")
summary(model_q10_1)
##
## Call:
## glm(formula = q10 ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + geo_region_in_cb_new + religion_in_new + q3_4,
## family = binomial, data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept 0.05288 0.11252 0.470 0.63840
## Partisan_Index -0.46352 0.02606 -17.785 < 2e-16 ***
## age 0.17325 0.02698 6.421 1.35e-10 ***
## Rest_Tiers -0.11421 0.04002 -2.854 0.00432 **
## BJP 0.13558 0.04963 2.732 0.00630 **
## INC -0.04090 0.07055 -0.580 0.56207
## SE -0.23159 0.04009 -5.776 7.63e-09 ***
## Hindu 0.31630 0.05497 5.754 8.74e-09 ***
## Muslim -0.34977 0.08057 -4.341 1.42e-05 ***
## Terrorism_Dissatisfied -0.10044 0.01905 -5.273 1.34e-07 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 16283 on 12433 degrees of freedom
## Residual deviance: 15326 on 12424 degrees of freedom
## AIC: 15230
##
## Number of Fisher Scoring iterations: 4
summ(model = model_q10_1, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
12434
|
|
Dependent variable
|
q10
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(9)
|
956.42
|
|
Pseudo-R² (Cragg-Uhler)
|
0.10
|
|
Pseudo-R² (McFadden)
|
0.06
|
|
AIC
|
15230.32
|
|
BIC
|
15304.60
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
1.05
|
0.85
|
1.31
|
0.47
|
0.64
|
|
Partisan_Index
|
0.63
|
0.60
|
0.66
|
-17.78
|
0.00
|
|
age
|
1.19
|
1.13
|
1.25
|
6.42
|
0.00
|
|
Rest_Tiers
|
0.89
|
0.82
|
0.96
|
-2.85
|
0.00
|
|
BJP
|
1.15
|
1.04
|
1.26
|
2.73
|
0.01
|
|
INC
|
0.96
|
0.84
|
1.10
|
-0.58
|
0.56
|
|
SE
|
0.79
|
0.73
|
0.86
|
-5.78
|
0.00
|
|
Hindu
|
1.37
|
1.23
|
1.53
|
5.75
|
0.00
|
|
Muslim
|
0.70
|
0.60
|
0.83
|
-4.34
|
0.00
|
|
Terrorism_Dissatisfied
|
0.90
|
0.87
|
0.94
|
-5.27
|
0.00
|
|
Standard errors: MLE
|
vif(model_q10_1)
## Warning in vif.default(model_q10_1): No intercept: vifs may not be sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 6.784445 1 2.604697
## age_cb_new 11.094830 1 3.330890
## geo_citytier_in 2.613490 1 1.616629
## p_party_new 4.408958 2 1.449052
## geo_region_in_cb_new 1.817544 1 1.348163
## religion_in_new 7.443501 2 1.651750
## q3_4 4.797057 1 2.190218
## DV <- age + partisanship + city tier + bordering + religion + party + dissatisfied.
## DV <- age + partisanship + city tier + distance + religion + party + dissatisfied.
model_q10_3 <-mint11_df_1 %>% glm(q10 ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+hub_dist+religion_in_new+q3_4, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_q10_3$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","Distance","Hindu","Muslim","Terrorism_Dissatisfied")
summary(model_q10_3)
##
## Call:
## glm(formula = q10 ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + hub_dist + religion_in_new + q3_4, family = binomial,
## data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept 1.464e-01 1.167e-01 1.254 0.20984
## Partisan_Index -4.623e-01 2.606e-02 -17.741 < 2e-16 ***
## age 1.726e-01 2.698e-02 6.396 1.59e-10 ***
## Rest_Tiers -1.015e-01 4.017e-02 -2.526 0.01153 *
## BJP 1.356e-01 4.960e-02 2.734 0.00627 **
## INC -3.873e-02 7.057e-02 -0.549 0.58312
## Distance -2.123e-04 3.606e-05 -5.887 3.93e-09 ***
## Hindu 3.113e-01 5.500e-02 5.659 1.52e-08 ***
## Muslim -3.503e-01 8.058e-02 -4.348 1.37e-05 ***
## Terrorism_Dissatisfied -9.975e-02 1.905e-02 -5.236 1.64e-07 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 16283 on 12433 degrees of freedom
## Residual deviance: 15325 on 12424 degrees of freedom
## AIC: 15228
##
## Number of Fisher Scoring iterations: 4
summ(model = model_q10_3, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
12434
|
|
Dependent variable
|
q10
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(9)
|
957.70
|
|
Pseudo-R² (Cragg-Uhler)
|
0.10
|
|
Pseudo-R² (McFadden)
|
0.06
|
|
AIC
|
15227.53
|
|
BIC
|
15301.82
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
1.16
|
0.92
|
1.46
|
1.25
|
0.21
|
|
Partisan_Index
|
0.63
|
0.60
|
0.66
|
-17.74
|
0.00
|
|
age
|
1.19
|
1.13
|
1.25
|
6.40
|
0.00
|
|
Rest_Tiers
|
0.90
|
0.84
|
0.98
|
-2.53
|
0.01
|
|
BJP
|
1.15
|
1.04
|
1.26
|
2.73
|
0.01
|
|
INC
|
0.96
|
0.84
|
1.10
|
-0.55
|
0.58
|
|
Distance
|
1.00
|
1.00
|
1.00
|
-5.89
|
0.00
|
|
Hindu
|
1.37
|
1.23
|
1.52
|
5.66
|
0.00
|
|
Muslim
|
0.70
|
0.60
|
0.82
|
-4.35
|
0.00
|
|
Terrorism_Dissatisfied
|
0.91
|
0.87
|
0.94
|
-5.24
|
0.00
|
|
Standard errors: MLE
|
vif(model_q10_3)
## Warning in vif.default(model_q10_3): No intercept: vifs may not be sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 6.780030 1 2.603849
## age_cb_new 11.093632 1 3.330710
## geo_citytier_in 2.631937 1 1.622325
## p_party_new 4.406691 2 1.448866
## hub_dist 4.070893 1 2.017645
## religion_in_new 7.451834 2 1.652212
## q3_4 4.797720 1 2.190370
# Set 6: (Q11b < dissatisfied with terrorism)
## DV <- age + partisanship + city tier + region (4) + religion + party + dissatisfied.
## DV <- age + partisanship + city tier + region (2) + religion + party + dissatisfied.
## DV <- age + partisanship + city tier + bordering + religion + party + dissatisfied.
## DV <- age + partisanship + city tier + distance + religion + party + dissatisfied.
# Set 7: (Q11 < cricket)
## DV <- age + partisanship + city tier + region (4) + religion + party + cricket.
model_10c_1 <-mint11_df_1 %>% glm(q10 ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+geo_region_in_cb_new+religion_in_new+q7_5, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_10c_1$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","SE","Hindu","Muslim","Cricket")
summary(model_10c_1)
##
## Call:
## glm(formula = q10 ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + geo_region_in_cb_new + religion_in_new + q7_5,
## family = binomial, data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept -0.78631 0.11980 -6.563 5.26e-11 ***
## Partisan_Index -0.47241 0.02601 -18.160 < 2e-16 ***
## age 0.18422 0.02707 6.805 1.01e-11 ***
## Rest_Tiers -0.11681 0.04015 -2.909 0.003622 **
## BJP 0.16837 0.04863 3.462 0.000536 ***
## INC -0.06695 0.07060 -0.948 0.342987
## SE -0.21167 0.04027 -5.256 1.47e-07 ***
## Hindu 0.32293 0.05491 5.881 4.07e-09 ***
## Muslim -0.37922 0.08042 -4.716 2.41e-06 ***
## Cricket 0.40925 0.03913 10.458 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 16283 on 12433 degrees of freedom
## Residual deviance: 15245 on 12424 degrees of freedom
## AIC: 15157
##
## Number of Fisher Scoring iterations: 4
summ(model = model_10c_1, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
12434
|
|
Dependent variable
|
q10
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(9)
|
1037.90
|
|
Pseudo-R² (Cragg-Uhler)
|
0.11
|
|
Pseudo-R² (McFadden)
|
0.06
|
|
AIC
|
15156.57
|
|
BIC
|
15230.85
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.46
|
0.36
|
0.58
|
-6.56
|
0.00
|
|
Partisan_Index
|
0.62
|
0.59
|
0.66
|
-18.16
|
0.00
|
|
age
|
1.20
|
1.14
|
1.27
|
6.80
|
0.00
|
|
Rest_Tiers
|
0.89
|
0.82
|
0.96
|
-2.91
|
0.00
|
|
BJP
|
1.18
|
1.08
|
1.30
|
3.46
|
0.00
|
|
INC
|
0.94
|
0.81
|
1.07
|
-0.95
|
0.34
|
|
SE
|
0.81
|
0.75
|
0.88
|
-5.26
|
0.00
|
|
Hindu
|
1.38
|
1.24
|
1.54
|
5.88
|
0.00
|
|
Muslim
|
0.68
|
0.58
|
0.80
|
-4.72
|
0.00
|
|
Cricket
|
1.51
|
1.39
|
1.63
|
10.46
|
0.00
|
|
Standard errors: MLE
|
vif(model_10c_1)
## Warning in vif.default(model_10c_1): No intercept: vifs may not be sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 6.710081 1 2.590382
## age_cb_new 11.083415 1 3.329176
## geo_citytier_in 2.611301 1 1.615952
## p_party_new 4.096095 2 1.422632
## geo_region_in_cb_new 1.820680 1 1.349326
## religion_in_new 7.203118 2 1.638250
## q7_5 10.214372 1 3.195993
## q9_2 DV <- age + partisanship + city tier + region (4) + religion + party
model_10c_2 <-mint11_df_1 %>% glm(q10 ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+geo_region_in_cb+religion_in_new+q7_5, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_10c_2$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","South India", "East & Northeast India", "West India","Hindu","Muslim","Cricket")
summary(model_10c_2)
##
## Call:
## glm(formula = q10 ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + geo_region_in_cb + religion_in_new + q7_5,
## family = binomial, data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept -0.77393 0.12243 -6.321 2.59e-10 ***
## Partisan_Index -0.47184 0.02610 -18.076 < 2e-16 ***
## age 0.18835 0.02714 6.939 3.95e-12 ***
## Rest_Tiers -0.10994 0.04025 -2.731 0.0063 **
## BJP 0.14228 0.04887 2.911 0.0036 **
## INC -0.04208 0.07093 -0.593 0.5530
## South India -0.35619 0.05221 -6.822 8.98e-12 ***
## East & Northeast India 0.05884 0.05680 1.036 0.3003
## West India 0.08734 0.05365 1.628 0.1035
## Hindu 0.29326 0.05531 5.302 1.14e-07 ***
## Muslim -0.40441 0.08078 -5.006 5.55e-07 ***
## Cricket 0.39683 0.03923 10.115 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 16283 on 12433 degrees of freedom
## Residual deviance: 15198 on 12422 degrees of freedom
## AIC: 15125
##
## Number of Fisher Scoring iterations: 4
summ(model = model_10c_2, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
12434
|
|
Dependent variable
|
q10
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(11)
|
1085.21
|
|
Pseudo-R² (Cragg-Uhler)
|
0.11
|
|
Pseudo-R² (McFadden)
|
0.07
|
|
AIC
|
15124.98
|
|
BIC
|
15214.12
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.46
|
0.36
|
0.59
|
-6.32
|
0.00
|
|
Partisan_Index
|
0.62
|
0.59
|
0.66
|
-18.08
|
0.00
|
|
age
|
1.21
|
1.14
|
1.27
|
6.94
|
0.00
|
|
Rest_Tiers
|
0.90
|
0.83
|
0.97
|
-2.73
|
0.01
|
|
BJP
|
1.15
|
1.05
|
1.27
|
2.91
|
0.00
|
|
INC
|
0.96
|
0.83
|
1.10
|
-0.59
|
0.55
|
|
South India
|
0.70
|
0.63
|
0.78
|
-6.82
|
0.00
|
|
East & Northeast India
|
1.06
|
0.95
|
1.19
|
1.04
|
0.30
|
|
West India
|
1.09
|
0.98
|
1.21
|
1.63
|
0.10
|
|
Hindu
|
1.34
|
1.20
|
1.49
|
5.30
|
0.00
|
|
Muslim
|
0.67
|
0.57
|
0.78
|
-5.01
|
0.00
|
|
Cricket
|
1.49
|
1.38
|
1.61
|
10.11
|
0.00
|
|
Standard errors: MLE
|
vif(model_10c_2)
## Warning in vif.default(model_10c_2): No intercept: vifs may not be sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 6.730975 1 2.594412
## age_cb_new 11.096739 1 3.331177
## geo_citytier_in 2.608948 1 1.615224
## p_party_new 4.157647 2 1.427946
## geo_region_in_cb 2.928743 3 1.196135
## religion_in_new 7.295276 2 1.643465
## q7_5 10.232139 1 3.198771
## q9_3 DV <- age + partisanship + city tier + distance + religion + party.
model_q10c_3 <-mint11_df_1 %>% glm(q10 ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+hub_dist+religion_in_new+q7_5, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_q10c_3$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","Distance","Hindu","Muslim","Cricket")
summary(model_q10c_3)
##
## Call:
## glm(formula = q10 ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + hub_dist + religion_in_new + q7_5, family = binomial,
## data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept -6.960e-01 1.240e-01 -5.612 2.00e-08 ***
## Partisan_Index -4.712e-01 2.601e-02 -18.119 < 2e-16 ***
## age 1.835e-01 2.707e-02 6.780 1.20e-11 ***
## Rest_Tiers -1.046e-01 4.030e-02 -2.597 0.009413 **
## BJP 1.665e-01 4.862e-02 3.424 0.000617 ***
## INC -6.494e-02 7.062e-02 -0.920 0.357797
## Distance -1.994e-04 3.623e-05 -5.504 3.72e-08 ***
## Hindu 3.178e-01 5.494e-02 5.784 7.29e-09 ***
## Muslim -3.797e-01 8.043e-02 -4.720 2.35e-06 ***
## Cricket 4.109e-01 3.912e-02 10.504 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 16283 on 12433 degrees of freedom
## Residual deviance: 15242 on 12424 degrees of freedom
## AIC: 15152
##
## Number of Fisher Scoring iterations: 4
summ(model = model_q10c_3, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
12434
|
|
Dependent variable
|
q10
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(9)
|
1040.56
|
|
Pseudo-R² (Cragg-Uhler)
|
0.11
|
|
Pseudo-R² (McFadden)
|
0.06
|
|
AIC
|
15152.44
|
|
BIC
|
15226.73
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.50
|
0.39
|
0.64
|
-5.61
|
0.00
|
|
Partisan_Index
|
0.62
|
0.59
|
0.66
|
-18.12
|
0.00
|
|
age
|
1.20
|
1.14
|
1.27
|
6.78
|
0.00
|
|
Rest_Tiers
|
0.90
|
0.83
|
0.97
|
-2.60
|
0.01
|
|
BJP
|
1.18
|
1.07
|
1.30
|
3.42
|
0.00
|
|
INC
|
0.94
|
0.82
|
1.08
|
-0.92
|
0.36
|
|
Distance
|
1.00
|
1.00
|
1.00
|
-5.50
|
0.00
|
|
Hindu
|
1.37
|
1.23
|
1.53
|
5.78
|
0.00
|
|
Muslim
|
0.68
|
0.58
|
0.80
|
-4.72
|
0.00
|
|
Cricket
|
1.51
|
1.40
|
1.63
|
10.50
|
0.00
|
|
Standard errors: MLE
|
vif(model_q10c_3)
## Warning in vif.default(model_q10c_3): No intercept: vifs may not be sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 6.705824 1 2.589561
## age_cb_new 11.082580 1 3.329051
## geo_citytier_in 2.629426 1 1.621551
## p_party_new 4.097809 2 1.422781
## hub_dist 4.078403 1 2.019506
## religion_in_new 7.209944 2 1.638638
## q7_5 10.207034 1 3.194845
## q9_4 DV <- age + partisanship + city tier + bordering + religion + party
model_10c_4 <-mint11_df_1 %>% glm(q10 ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+bordering+religion_in_new+q7_5, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_10c_4$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","bordering","Hindu","Muslim","Cricket")
summary(model_10c_4)
##
## Call:
## glm(formula = q10 ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + bordering + religion_in_new + q7_5, family = binomial,
## data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept -0.80198 0.18707 -4.287 1.81e-05 ***
## Partisan_Index -0.46918 0.02596 -18.071 < 2e-16 ***
## age 0.18609 0.02705 6.879 6.02e-12 ***
## Rest_Tiers -0.13522 0.04130 -3.274 0.00106 **
## BJP 0.21116 0.04790 4.408 1.04e-05 ***
## INC -0.06728 0.07046 -0.955 0.33968
## bordering -0.06064 0.07183 -0.844 0.39853
## Hindu 0.32787 0.05485 5.977 2.27e-09 ***
## Muslim -0.36883 0.08031 -4.593 4.37e-06 ***
## Cricket 0.41935 0.03905 10.739 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 16283 on 12433 degrees of freedom
## Residual deviance: 15272 on 12424 degrees of freedom
## AIC: 15186
##
## Number of Fisher Scoring iterations: 4
summ(model = model_10c_4, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
12434
|
|
Dependent variable
|
q10
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(9)
|
1010.95
|
|
Pseudo-R² (Cragg-Uhler)
|
0.11
|
|
Pseudo-R² (McFadden)
|
0.06
|
|
AIC
|
15186.05
|
|
BIC
|
15260.34
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.45
|
0.31
|
0.65
|
-4.29
|
0.00
|
|
Partisan_Index
|
0.63
|
0.59
|
0.66
|
-18.07
|
0.00
|
|
age
|
1.20
|
1.14
|
1.27
|
6.88
|
0.00
|
|
Rest_Tiers
|
0.87
|
0.81
|
0.95
|
-3.27
|
0.00
|
|
BJP
|
1.24
|
1.12
|
1.36
|
4.41
|
0.00
|
|
INC
|
0.93
|
0.81
|
1.07
|
-0.95
|
0.34
|
|
bordering
|
0.94
|
0.82
|
1.08
|
-0.84
|
0.40
|
|
Hindu
|
1.39
|
1.25
|
1.55
|
5.98
|
0.00
|
|
Muslim
|
0.69
|
0.59
|
0.81
|
-4.59
|
0.00
|
|
Cricket
|
1.52
|
1.41
|
1.64
|
10.74
|
0.00
|
|
Standard errors: MLE
|
vif(model_10c_4)
## Warning in vif.default(model_10c_4): No intercept: vifs may not be sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 6.696361 1 2.587733
## age_cb_new 11.095856 1 3.331044
## geo_citytier_in 2.771928 1 1.664911
## p_party_new 3.970333 2 1.411584
## bordering 51.321970 1 7.163935
## religion_in_new 7.204149 2 1.638308
## q7_5 10.192069 1 3.192502
# Set 8: (Q11 < restrain)
##DV <- age + partisanship + city tier + region (4) + religion + party + restrain.
#DV <- age + partisanship + city tier + region (2) + religion + party + restrain.
#DV <- age + partisanship + city tier + bordering + religion + party + restrain.
#DV <- age + partisanship + city tier + distance + religion + party + restrain.
model_10r_1 <-mint11_df_1 %>% glm(q10 ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+geo_region_in_cb_new+religion_in_new+q7_1, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_10r_1$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","SE","Hindu","Muslim","Restrain")
summary(model_10r_1)
##
## Call:
## glm(formula = q10 ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + geo_region_in_cb_new + religion_in_new + q7_1,
## family = binomial, data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept -0.42748 0.11876 -3.599 0.000319 ***
## Partisan_Index -0.48302 0.02593 -18.627 < 2e-16 ***
## age 0.18265 0.02698 6.771 1.28e-11 ***
## Rest_Tiers -0.11369 0.04000 -2.842 0.004483 **
## BJP 0.19545 0.04837 4.040 5.34e-05 ***
## INC -0.06753 0.07025 -0.961 0.336416
## SE -0.22973 0.04007 -5.733 9.87e-09 ***
## Hindu 0.34566 0.05468 6.322 2.59e-10 ***
## Muslim -0.39034 0.08012 -4.872 1.10e-06 ***
## Restrain 0.17202 0.03889 4.423 9.71e-06 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 16283 on 12433 degrees of freedom
## Residual deviance: 15335 on 12424 degrees of freedom
## AIC: 15240
##
## Number of Fisher Scoring iterations: 4
summ(model = model_10r_1, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
12434
|
|
Dependent variable
|
q10
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(9)
|
947.91
|
|
Pseudo-R² (Cragg-Uhler)
|
0.10
|
|
Pseudo-R² (McFadden)
|
0.06
|
|
AIC
|
15240.47
|
|
BIC
|
15314.75
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.65
|
0.52
|
0.82
|
-3.60
|
0.00
|
|
Partisan_Index
|
0.62
|
0.59
|
0.65
|
-18.63
|
0.00
|
|
age
|
1.20
|
1.14
|
1.27
|
6.77
|
0.00
|
|
Rest_Tiers
|
0.89
|
0.83
|
0.97
|
-2.84
|
0.00
|
|
BJP
|
1.22
|
1.11
|
1.34
|
4.04
|
0.00
|
|
INC
|
0.93
|
0.81
|
1.07
|
-0.96
|
0.34
|
|
SE
|
0.79
|
0.73
|
0.86
|
-5.73
|
0.00
|
|
Hindu
|
1.41
|
1.27
|
1.57
|
6.32
|
0.00
|
|
Muslim
|
0.68
|
0.58
|
0.79
|
-4.87
|
0.00
|
|
Restrain
|
1.19
|
1.10
|
1.28
|
4.42
|
0.00
|
|
Standard errors: MLE
|
vif(model_10r_1)
## Warning in vif.default(model_10r_1): No intercept: vifs may not be sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 6.717425 1 2.591800
## age_cb_new 11.089832 1 3.330140
## geo_citytier_in 2.611485 1 1.616009
## p_party_new 4.078217 2 1.421077
## geo_region_in_cb_new 1.818336 1 1.348457
## religion_in_new 7.186341 2 1.637295
## q7_1 9.659471 1 3.107969
#q9_2 DV <- age + partisanship + city tier + region (4) + religion + party
model_10r_2 <-mint11_df_1 %>% glm(q10 ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+geo_region_in_cb+religion_in_new+q7_1, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_10r_2$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","South India", "East & Northeast India", "West India","Hindu","Muslim","Restraint")
summary(model_10r_2)
##
## Call:
## glm(formula = q10 ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + geo_region_in_cb + religion_in_new + q7_1,
## family = binomial, data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept -0.42036 0.12167 -3.455 0.000550 ***
## Partisan_Index -0.48204 0.02603 -18.522 < 2e-16 ***
## age 0.18707 0.02706 6.914 4.72e-12 ***
## Rest_Tiers -0.10635 0.04011 -2.652 0.008008 **
## BJP 0.16738 0.04863 3.442 0.000578 ***
## INC -0.04138 0.07061 -0.586 0.557810
## South India -0.37808 0.05199 -7.273 3.52e-13 ***
## East & Northeast India 0.05387 0.05657 0.952 0.340909
## West India 0.09533 0.05348 1.782 0.074674 .
## Hindu 0.31431 0.05509 5.706 1.16e-08 ***
## Muslim -0.41592 0.08049 -5.168 2.37e-07 ***
## Restraint 0.16129 0.03900 4.135 3.54e-05 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 16283 on 12433 degrees of freedom
## Residual deviance: 15283 on 12422 degrees of freedom
## AIC: 15205
##
## Number of Fisher Scoring iterations: 4
summ(model = model_10r_2, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
12434
|
|
Dependent variable
|
q10
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(11)
|
999.87
|
|
Pseudo-R² (Cragg-Uhler)
|
0.10
|
|
Pseudo-R² (McFadden)
|
0.06
|
|
AIC
|
15205.11
|
|
BIC
|
15294.25
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.66
|
0.52
|
0.83
|
-3.46
|
0.00
|
|
Partisan_Index
|
0.62
|
0.59
|
0.65
|
-18.52
|
0.00
|
|
age
|
1.21
|
1.14
|
1.27
|
6.91
|
0.00
|
|
Rest_Tiers
|
0.90
|
0.83
|
0.97
|
-2.65
|
0.01
|
|
BJP
|
1.18
|
1.07
|
1.30
|
3.44
|
0.00
|
|
INC
|
0.96
|
0.84
|
1.10
|
-0.59
|
0.56
|
|
South India
|
0.69
|
0.62
|
0.76
|
-7.27
|
0.00
|
|
East & Northeast India
|
1.06
|
0.94
|
1.18
|
0.95
|
0.34
|
|
West India
|
1.10
|
0.99
|
1.22
|
1.78
|
0.07
|
|
Hindu
|
1.37
|
1.23
|
1.53
|
5.71
|
0.00
|
|
Muslim
|
0.66
|
0.56
|
0.77
|
-5.17
|
0.00
|
|
Restraint
|
1.18
|
1.09
|
1.27
|
4.14
|
0.00
|
|
Standard errors: MLE
|
vif(model_10r_2)
## Warning in vif.default(model_10r_2): No intercept: vifs may not be sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 6.738170 1 2.595798
## age_cb_new 11.106370 1 3.332622
## geo_citytier_in 2.609164 1 1.615291
## p_party_new 4.141460 2 1.426554
## geo_region_in_cb 2.923550 3 1.195781
## religion_in_new 7.278213 2 1.642503
## q7_1 9.678312 1 3.110999
#q9_3 DV <- age + partisanship + city tier + distance + religion + party.
model_q10r_3 <-mint11_df_1 %>% glm(q10 ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+hub_dist+religion_in_new+q7_1, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_q10r_3$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","Distance","Hindu","Muslim","Restraint")
summary(model_q10r_3)
##
## Call:
## glm(formula = q10 ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + hub_dist + religion_in_new + q7_1, family = binomial,
## data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept -3.306e-01 1.230e-01 -2.688 0.00718 **
## Partisan_Index -4.817e-01 2.592e-02 -18.581 < 2e-16 ***
## age 1.818e-01 2.697e-02 6.741 1.57e-11 ***
## Rest_Tiers -1.010e-01 4.015e-02 -2.515 0.01189 *
## BJP 1.948e-01 4.836e-02 4.028 5.63e-05 ***
## INC -6.535e-02 7.027e-02 -0.930 0.35236
## Distance -2.110e-04 3.604e-05 -5.854 4.81e-09 ***
## Hindu 3.405e-01 5.471e-02 6.225 4.82e-10 ***
## Muslim -3.907e-01 8.013e-02 -4.875 1.09e-06 ***
## Restraint 1.708e-01 3.889e-02 4.391 1.13e-05 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 16283 on 12433 degrees of freedom
## Residual deviance: 15334 on 12424 degrees of freedom
## AIC: 15237
##
## Number of Fisher Scoring iterations: 4
summ(model = model_q10r_3, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
12434
|
|
Dependent variable
|
q10
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(9)
|
949.30
|
|
Pseudo-R² (Cragg-Uhler)
|
0.10
|
|
Pseudo-R² (McFadden)
|
0.06
|
|
AIC
|
15237.42
|
|
BIC
|
15311.70
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.72
|
0.56
|
0.91
|
-2.69
|
0.01
|
|
Partisan_Index
|
0.62
|
0.59
|
0.65
|
-18.58
|
0.00
|
|
age
|
1.20
|
1.14
|
1.26
|
6.74
|
0.00
|
|
Rest_Tiers
|
0.90
|
0.84
|
0.98
|
-2.52
|
0.01
|
|
BJP
|
1.22
|
1.11
|
1.34
|
4.03
|
0.00
|
|
INC
|
0.94
|
0.82
|
1.08
|
-0.93
|
0.35
|
|
Distance
|
1.00
|
1.00
|
1.00
|
-5.85
|
0.00
|
|
Hindu
|
1.41
|
1.26
|
1.56
|
6.22
|
0.00
|
|
Muslim
|
0.68
|
0.58
|
0.79
|
-4.88
|
0.00
|
|
Restraint
|
1.19
|
1.10
|
1.28
|
4.39
|
0.00
|
|
Standard errors: MLE
|
vif(model_q10r_3)
## Warning in vif.default(model_q10r_3): No intercept: vifs may not be sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 6.712870 1 2.590921
## age_cb_new 11.088546 1 3.329947
## geo_citytier_in 2.629877 1 1.621689
## p_party_new 4.079264 2 1.421168
## hub_dist 4.072915 1 2.018146
## religion_in_new 7.193276 2 1.637690
## q7_1 9.661287 1 3.108261
#q9_4 DV <- age + partisanship + city tier + bordering + religion + party
model_10r_4 <-mint11_df_1 %>% glm(q10 ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+bordering+religion_in_new+q7_1, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_10r_4$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","bordering","Hindu","Muslim","Restraint")
summary(model_10r_4)
##
## Call:
## glm(formula = q10 ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + bordering + religion_in_new + q7_1, family = binomial,
## data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept -0.44792 0.18670 -2.399 0.01644 *
## Partisan_Index -0.47981 0.02588 -18.543 < 2e-16 ***
## age 0.18444 0.02695 6.844 7.70e-12 ***
## Rest_Tiers -0.13266 0.04113 -3.225 0.00126 **
## BJP 0.24282 0.04762 5.099 3.42e-07 ***
## INC -0.06798 0.07011 -0.970 0.33222
## bordering -0.05808 0.07161 -0.811 0.41732
## Hindu 0.35199 0.05461 6.445 1.15e-10 ***
## Muslim -0.37899 0.07999 -4.738 2.16e-06 ***
## Restraint 0.17483 0.03883 4.502 6.72e-06 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 16283 on 12433 degrees of freedom
## Residual deviance: 15367 on 12424 degrees of freedom
## AIC: 15276
##
## Number of Fisher Scoring iterations: 4
summ(model = model_10r_4, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
12434
|
|
Dependent variable
|
q10
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(9)
|
915.64
|
|
Pseudo-R² (Cragg-Uhler)
|
0.10
|
|
Pseudo-R² (McFadden)
|
0.06
|
|
AIC
|
15275.68
|
|
BIC
|
15349.96
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.64
|
0.44
|
0.92
|
-2.40
|
0.02
|
|
Partisan_Index
|
0.62
|
0.59
|
0.65
|
-18.54
|
0.00
|
|
age
|
1.20
|
1.14
|
1.27
|
6.84
|
0.00
|
|
Rest_Tiers
|
0.88
|
0.81
|
0.95
|
-3.23
|
0.00
|
|
BJP
|
1.27
|
1.16
|
1.40
|
5.10
|
0.00
|
|
INC
|
0.93
|
0.81
|
1.07
|
-0.97
|
0.33
|
|
bordering
|
0.94
|
0.82
|
1.09
|
-0.81
|
0.42
|
|
Hindu
|
1.42
|
1.28
|
1.58
|
6.45
|
0.00
|
|
Muslim
|
0.68
|
0.59
|
0.80
|
-4.74
|
0.00
|
|
Restraint
|
1.19
|
1.10
|
1.29
|
4.50
|
0.00
|
|
Standard errors: MLE
|
vif(model_10r_4)
## Warning in vif.default(model_10r_4): No intercept: vifs may not be sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 6.702512 1 2.588921
## age_cb_new 11.103402 1 3.332177
## geo_citytier_in 2.771898 1 1.664902
## p_party_new 3.950403 2 1.409809
## bordering 51.415419 1 7.170455
## religion_in_new 7.189259 2 1.637461
## q7_1 9.658804 1 3.107862
#Set 9: robustness, caste (Q11 < caste)
#DV <- age + partisanship + city tier + region (4) + religion + party + caste.
#DV <- age + partisanship + city tier + region (2) + religion + party + caste.
#DV <- age + partisanship + city tier + bordering + religion + party + caste.
#DV <- age + partisanship + city tier + distance + religion + party + caste.
model_10cas_1 <-mint11_df_1 %>% glm(q10 ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+geo_region_in_cb_new+religion_in_new+caste_in_new, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_10cas_1$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","SE","Hindu","Muslim","Caste")
summary(model_10cas_1)
##
## Call:
## glm(formula = q10 ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + geo_region_in_cb_new + religion_in_new + caste_in_new,
## family = binomial, data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept -0.15524 0.10789 -1.439 0.15018
## Partisan_Index -0.47829 0.02590 -18.468 < 2e-16 ***
## age 0.18016 0.02712 6.642 3.10e-11 ***
## Rest_Tiers -0.11040 0.04003 -2.758 0.00581 **
## BJP 0.19123 0.04862 3.933 8.38e-05 ***
## INC -0.07197 0.07022 -1.025 0.30535
## SE -0.23026 0.04017 -5.732 9.90e-09 ***
## Hindu 0.34237 0.05502 6.223 4.88e-10 ***
## Muslim -0.39543 0.08035 -4.921 8.60e-07 ***
## Caste -0.01126 0.01800 -0.626 0.53159
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 16283 on 12433 degrees of freedom
## Residual deviance: 15354 on 12424 degrees of freedom
## AIC: 15261
##
## Number of Fisher Scoring iterations: 4
summ(model = model_10cas_1, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
12434
|
|
Dependent variable
|
q10
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(9)
|
928.72
|
|
Pseudo-R² (Cragg-Uhler)
|
0.10
|
|
Pseudo-R² (McFadden)
|
0.06
|
|
AIC
|
15261.29
|
|
BIC
|
15335.57
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.86
|
0.69
|
1.06
|
-1.44
|
0.15
|
|
Partisan_Index
|
0.62
|
0.59
|
0.65
|
-18.47
|
0.00
|
|
age
|
1.20
|
1.14
|
1.26
|
6.64
|
0.00
|
|
Rest_Tiers
|
0.90
|
0.83
|
0.97
|
-2.76
|
0.01
|
|
BJP
|
1.21
|
1.10
|
1.33
|
3.93
|
0.00
|
|
INC
|
0.93
|
0.81
|
1.07
|
-1.03
|
0.31
|
|
SE
|
0.79
|
0.73
|
0.86
|
-5.73
|
0.00
|
|
Hindu
|
1.41
|
1.26
|
1.57
|
6.22
|
0.00
|
|
Muslim
|
0.67
|
0.58
|
0.79
|
-4.92
|
0.00
|
|
Caste
|
0.99
|
0.95
|
1.02
|
-0.63
|
0.53
|
|
Standard errors: MLE
|
vif(model_10cas_1)
## Warning in vif.default(model_10cas_1): No intercept: vifs may not be sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 6.710504 1 2.590464
## age_cb_new 11.232543 1 3.351499
## geo_citytier_in 2.620174 1 1.618695
## p_party_new 4.125800 2 1.425204
## geo_region_in_cb_new 1.829482 1 1.352584
## religion_in_new 7.291667 2 1.643262
## caste_in_new 3.795605 1 1.948231
#q9_2 DV <- age + partisanship + city tier + region (4) + religion + party
model_10cas_2 <-mint11_df_1 %>% glm(q10 ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+geo_region_in_cb+religion_in_new+caste_in_new, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_10cas_2$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","South India", "East & Northeast India", "West India","Hindu","Muslim","Caste")
summary(model_10cas_2)
##
## Call:
## glm(formula = q10 ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + geo_region_in_cb + religion_in_new + caste_in_new,
## family = binomial, data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept -0.168895 0.110721 -1.525 0.127157
## Partisan_Index -0.477559 0.025994 -18.372 < 2e-16 ***
## age 0.184103 0.027210 6.766 1.32e-11 ***
## Rest_Tiers -0.103561 0.040139 -2.580 0.009879 **
## BJP 0.163717 0.048865 3.350 0.000807 ***
## INC -0.044632 0.070586 -0.632 0.527183
## South India -0.384265 0.052089 -7.377 1.62e-13 ***
## East & Northeast India 0.057875 0.056568 1.023 0.306257
## West India 0.092761 0.053431 1.736 0.082546 .
## Hindu 0.311602 0.055427 5.622 1.89e-08 ***
## Muslim -0.420229 0.080752 -5.204 1.95e-07 ***
## Caste -0.007191 0.018045 -0.399 0.690244
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 16283 on 12433 degrees of freedom
## Residual deviance: 15300 on 12422 degrees of freedom
## AIC: 15224
##
## Number of Fisher Scoring iterations: 4
summ(model = model_10cas_2, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
12434
|
|
Dependent variable
|
q10
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(11)
|
982.92
|
|
Pseudo-R² (Cragg-Uhler)
|
0.10
|
|
Pseudo-R² (McFadden)
|
0.06
|
|
AIC
|
15223.92
|
|
BIC
|
15313.06
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.84
|
0.68
|
1.05
|
-1.53
|
0.13
|
|
Partisan_Index
|
0.62
|
0.59
|
0.65
|
-18.37
|
0.00
|
|
age
|
1.20
|
1.14
|
1.27
|
6.77
|
0.00
|
|
Rest_Tiers
|
0.90
|
0.83
|
0.98
|
-2.58
|
0.01
|
|
BJP
|
1.18
|
1.07
|
1.30
|
3.35
|
0.00
|
|
INC
|
0.96
|
0.83
|
1.10
|
-0.63
|
0.53
|
|
South India
|
0.68
|
0.61
|
0.75
|
-7.38
|
0.00
|
|
East & Northeast India
|
1.06
|
0.95
|
1.18
|
1.02
|
0.31
|
|
West India
|
1.10
|
0.99
|
1.22
|
1.74
|
0.08
|
|
Hindu
|
1.37
|
1.23
|
1.52
|
5.62
|
0.00
|
|
Muslim
|
0.66
|
0.56
|
0.77
|
-5.20
|
0.00
|
|
Caste
|
0.99
|
0.96
|
1.03
|
-0.40
|
0.69
|
|
Standard errors: MLE
|
vif(model_10cas_2)
## Warning in vif.default(model_10cas_2): No intercept: vifs may not be sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 6.731195 1 2.594455
## age_cb_new 11.250455 1 3.354170
## geo_citytier_in 2.618440 1 1.618159
## p_party_new 4.187163 2 1.430474
## geo_region_in_cb 2.938522 3 1.196800
## religion_in_new 7.385025 2 1.648496
## caste_in_new 3.794444 1 1.947933
#q9_3 DV <- age + partisanship + city tier + distance + religion + party.
model_q10cas_3 <-mint11_df_1 %>% glm(q10 ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+hub_dist+religion_in_new+caste_in_new, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_q10cas_3$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","Distance","Hindu","Muslim","Caste")
summary(model_q10cas_3)
##
## Call:
## glm(formula = q10 ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + hub_dist + religion_in_new + caste_in_new,
## family = binomial, data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept -6.013e-02 1.120e-01 -0.537 0.591
## Partisan_Index -4.770e-01 2.589e-02 -18.424 < 2e-16 ***
## age 1.793e-01 2.712e-02 6.611 3.82e-11 ***
## Rest_Tiers -9.767e-02 4.017e-02 -2.432 0.015 *
## BJP 1.906e-01 4.860e-02 3.920 8.84e-05 ***
## INC -6.973e-02 7.023e-02 -0.993 0.321
## Distance -2.123e-04 3.613e-05 -5.875 4.24e-09 ***
## Hindu 3.374e-01 5.504e-02 6.129 8.83e-10 ***
## Muslim -3.956e-01 8.036e-02 -4.923 8.53e-07 ***
## Caste -1.075e-02 1.800e-02 -0.597 0.550
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 16283 on 12433 degrees of freedom
## Residual deviance: 15352 on 12424 degrees of freedom
## AIC: 15258
##
## Number of Fisher Scoring iterations: 4
summ(model = model_q10cas_3, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
12434
|
|
Dependent variable
|
q10
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(9)
|
930.37
|
|
Pseudo-R² (Cragg-Uhler)
|
0.10
|
|
Pseudo-R² (McFadden)
|
0.06
|
|
AIC
|
15257.98
|
|
BIC
|
15332.26
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.94
|
0.76
|
1.17
|
-0.54
|
0.59
|
|
Partisan_Index
|
0.62
|
0.59
|
0.65
|
-18.42
|
0.00
|
|
age
|
1.20
|
1.13
|
1.26
|
6.61
|
0.00
|
|
Rest_Tiers
|
0.91
|
0.84
|
0.98
|
-2.43
|
0.02
|
|
BJP
|
1.21
|
1.10
|
1.33
|
3.92
|
0.00
|
|
INC
|
0.93
|
0.81
|
1.07
|
-0.99
|
0.32
|
|
Distance
|
1.00
|
1.00
|
1.00
|
-5.87
|
0.00
|
|
Hindu
|
1.40
|
1.26
|
1.56
|
6.13
|
0.00
|
|
Muslim
|
0.67
|
0.58
|
0.79
|
-4.92
|
0.00
|
|
Caste
|
0.99
|
0.96
|
1.02
|
-0.60
|
0.55
|
|
Standard errors: MLE
|
vif(model_q10cas_3)
## Warning in vif.default(model_q10cas_3): No intercept: vifs may not be sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 6.705690 1 2.589535
## age_cb_new 11.232633 1 3.351512
## geo_citytier_in 2.637213 1 1.623950
## p_party_new 4.126523 2 1.425267
## hub_dist 4.099425 1 2.024704
## religion_in_new 7.296794 2 1.643550
## caste_in_new 3.796486 1 1.948457
#q9_4 DV <- age + partisanship + city tier + bordering + religion + party
model_10cas_4 <-mint11_df_1 %>% glm(q10 ~ partisan_index+age_cb_new+ geo_citytier_in+ p_party_new+bordering+religion_in_new+caste_in_new, data=., family = binomial,weight)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
names(model_10cas_4$coefficients) <- c("intercept","Partisan_Index","age","Rest_Tiers","BJP","INC","bordering","Hindu","Muslim","Caste")
summary(model_10cas_4)
##
## Call:
## glm(formula = q10 ~ partisan_index + age_cb_new + geo_citytier_in +
## p_party_new + bordering + religion_in_new + caste_in_new,
## family = binomial, data = ., weights = weight)
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## intercept -0.15464 0.17951 -0.861 0.38901
## Partisan_Index -0.47470 0.02584 -18.371 < 2e-16 ***
## age 0.18340 0.02710 6.768 1.30e-11 ***
## Rest_Tiers -0.12872 0.04117 -3.127 0.00177 **
## BJP 0.23627 0.04793 4.929 8.26e-07 ***
## INC -0.07346 0.07007 -1.048 0.29441
## bordering -0.05991 0.07152 -0.838 0.40222
## Hindu 0.34567 0.05496 6.290 3.18e-10 ***
## Muslim -0.38687 0.08022 -4.823 1.42e-06 ***
## Caste -0.01921 0.01796 -1.070 0.28470
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 16283 on 12433 degrees of freedom
## Residual deviance: 15386 on 12424 degrees of freedom
## AIC: 15297
##
## Number of Fisher Scoring iterations: 4
summ(model = model_10cas_4, exp = TRUE)
## Warning in eval(family$initialize): non-integer #successes in a binomial glm!
|
Observations
|
12434
|
|
Dependent variable
|
q10
|
|
Type
|
Generalized linear model
|
|
Family
|
binomial
|
|
Link
|
logit
|
|
𝛘²(9)
|
896.50
|
|
Pseudo-R² (Cragg-Uhler)
|
0.09
|
|
Pseudo-R² (McFadden)
|
0.05
|
|
AIC
|
15296.68
|
|
BIC
|
15370.97
|
|
|
exp(Est.)
|
2.5%
|
97.5%
|
z val.
|
p
|
|
intercept
|
0.86
|
0.60
|
1.22
|
-0.86
|
0.39
|
|
Partisan_Index
|
0.62
|
0.59
|
0.65
|
-18.37
|
0.00
|
|
age
|
1.20
|
1.14
|
1.27
|
6.77
|
0.00
|
|
Rest_Tiers
|
0.88
|
0.81
|
0.95
|
-3.13
|
0.00
|
|
BJP
|
1.27
|
1.15
|
1.39
|
4.93
|
0.00
|
|
INC
|
0.93
|
0.81
|
1.07
|
-1.05
|
0.29
|
|
bordering
|
0.94
|
0.82
|
1.08
|
-0.84
|
0.40
|
|
Hindu
|
1.41
|
1.27
|
1.57
|
6.29
|
0.00
|
|
Muslim
|
0.68
|
0.58
|
0.79
|
-4.82
|
0.00
|
|
Caste
|
0.98
|
0.95
|
1.02
|
-1.07
|
0.28
|
|
Standard errors: MLE
|
vif(model_10cas_4)
## Warning in vif.default(model_10cas_4): No intercept: vifs may not be sensible.
## GVIF Df GVIF^(1/(2*Df))
## partisan_index 6.693905 1 2.587258
## age_cb_new 11.244785 1 3.353324
## geo_citytier_in 2.782756 1 1.668160
## p_party_new 4.008663 2 1.414979
## bordering 51.374073 1 7.167571
## religion_in_new 7.296877 2 1.643555
## caste_in_new 3.783423 1 1.945102