library(haven)
library(foreign) 
library(readr)
library(dplyr)

## 
## Attaching package: 'dplyr'

## The following objects are masked from 'package:stats':
## 
##     filter, lag

## The following objects are masked from 'package:base':
## 
##     intersect, setdiff, setequal, union

library(ggplot2)
library(broom)
library(car)

## Loading required package: carData

## 
## Attaching package: 'car'

## The following object is masked from 'package:dplyr':
## 
##     recode

library(readr) 
library(MASS)

## 
## Attaching package: 'MASS'

## The following object is masked from 'package:dplyr':
## 
##     select

library(car) 
library(lmtest)

## Loading required package: zoo

## 
## Attaching package: 'zoo'

## The following objects are masked from 'package:base':
## 
##     as.Date, as.Date.numeric

library(alr3)

## 
## Attaching package: 'alr3'

## The following object is masked from 'package:MASS':
## 
##     forbes

library(zoo)
library(nortest)
library(plotrix)
library(scales)

## 
## Attaching package: 'scales'

## The following object is masked from 'package:plotrix':
## 
##     rescale

## The following object is masked from 'package:readr':
## 
##     col_factor

library(tableone)
library(Weighted.Desc.Stat)
library(mitools)
library(survey)

## Loading required package: grid

## Loading required package: Matrix

## Loading required package: survival

## 
## Attaching package: 'survey'

## The following object is masked from 'package:graphics':
## 
##     dotchart

library(VGAM)

## Loading required package: stats4

## Loading required package: splines

## 
## Attaching package: 'VGAM'

## The following object is masked from 'package:survey':
## 
##     calibrate

## The following object is masked from 'package:lmtest':
## 
##     lrtest

## The following object is masked from 'package:car':
## 
##     logit

library(stargazer)

## 
## Please cite as:

##  Hlavac, Marek (2018). stargazer: Well-Formatted Regression and Summary Statistics Tables.

##  R package version 5.2.2. https://CRAN.R-project.org/package=stargazer

anes2016<-read_dta("C:\\Users\\Jaire\\OneDrive\\Desktop\\Stats for Dem Data 2\\Homework 3\\ANES2016.dta")

Outcome of Interest

In this analysis, the multinomial outome of interest is political party. The variable “V161155” within the ANES 2016 cross-sectional data set contains responses for the self-identified political party of respondents.

The survey question: “Generally speaking, do you usually think of yourself as a Democrat, a Republican, an independent, or what?”

# Outcome Raw Data:
table(anes2016$V161155)

## 
##   -9   -8    0    1    2    3    5 
##   15   10   49 1451 1231 1367  148

# Outcome Transformation:
anes2016$PolParty<-recode(anes2016$V161155, recodes = "-9:0=NA; 1='Democrat'; 2='Republican'; 3='Independent'; 5=NA", as.factor = T)

# Political Party from V161155

table(anes2016$PolParty)

## 
##    Democrat Independent  Republican 
##        1451        1367        1231

Research Question:

To what extent does self-identified race and highest level of education impact party preference?

Predictors:

Self-Identified Race -

# Raw data:
table(anes2016$V161310X)

## 
##   -9    1    2    3    4    5    6 
##   33 3038  398  148   27  450  177

# Transformation:
anes2016$Race<-recode(anes2016$V161310X, recodes = "-9=NA; 1='White'; 2='Black'; 3='AsNatPac'; 4='NatAmAlsk'; 5='Hisp'; 6='Other'", as.factor = T)

# Race from anes2016$V161310X
table(anes2016$Race)

## 
##  AsNatPac     Black      Hisp NatAmAlsk     Other     White 
##       148       398       450        27       177      3038

Highest Level of Education -

# Highest Level of Education Raw Data:

table(anes2016$V161270)

## 
##  -9   1   2   3   4   5   6   7   8   9  10  11  12  13  14  15  16  90  95 
##  15   1   3  15  22  32  40  62 107 810 899 313 288 955 499  88  93   5  24

# Highest Level of Education Transformations:
anes2016$HLEducation<-recode(anes2016$V161270, recodes = "15:95=NA;-9=NA;10=NA;1:8='LessHSDipGED'; 9='HSDipGED'; 11:12='Assoc'; 13='Bachlrs'; 14='Mastrs'", as.factor = T)

# Highest Level of Education from V161270:

table(anes2016$HLEducation)

## 
##        Assoc      Bachlrs     HSDipGED LessHSDipGED       Mastrs 
##          601          955          810          282          499

Analysis:

Proportional Odds Model:

The basic assumption of a proportional odds model is that for each level within the outcome variable (i.e. Politcal Party) other than the reference category (i.e. independent), the predictors are independent of the remaining outcome categories. However, this anaylsis uses a multinomial outcome rather than an ordinal outcome and uses a multinomial distribution.

Survey weight -

# Combining pre- post- weights into one object:

anes2016$fullwt<-as.numeric(anes2016$V160102+anes2016$V160101)

Multinomial Model -

mfit<-vglm(anes2016$PolParty~anes2016$HLEducation+anes2016$Race,
           family = multinomial(refLevel = 3), 
           data = anes2016, 
           weights =anes2016$fullwt/mean(anes2016$fullwt, na.rm=T))
summary(mfit)

## 
## Call:
## vglm(formula = anes2016$PolParty ~ anes2016$HLEducation + anes2016$Race, 
##     family = multinomial(refLevel = 3), data = anes2016, weights = anes2016$fullwt/mean(anes2016$fullwt, 
##         na.rm = T))
## 
## Pearson residuals:
##                       Min      1Q  Median     3Q   Max
## log(mu[,1]/mu[,3]) -6.475 -0.6900 -0.2980 0.8177 3.528
## log(mu[,2]/mu[,3]) -4.971 -0.6899 -0.3228 0.9798 4.605
## 
## Coefficients: 
##                                      Estimate Std. Error z value Pr(>|z|)    
## (Intercept):1                       0.5614827  0.3188524   1.761  0.07825 .  
## (Intercept):2                       0.6292030  0.3187755   1.974  0.04840 *  
## anes2016$HLEducationBachlrs:1      -0.0005426  0.1496642  -0.004  0.99711    
## anes2016$HLEducationBachlrs:2      -0.0612833  0.1452091  -0.422  0.67300    
## anes2016$HLEducationHSDipGED:1      0.1507730  0.1409454   1.070  0.28474    
## anes2016$HLEducationHSDipGED:2      0.1853885  0.1364104   1.359  0.17413    
## anes2016$HLEducationLessHSDipGED:1  0.2003232  0.1891991   1.059  0.28969    
## anes2016$HLEducationLessHSDipGED:2  0.3776572  0.1820974   2.074  0.03809 *  
## anes2016$HLEducationMastrs:1        0.9057196  0.1803981   5.021 5.15e-07 ***
## anes2016$HLEducationMastrs:2        0.3737575  0.1848012   2.022  0.04313 *  
## anes2016$RaceBlack:1                1.7726251  0.3768979   4.703 2.56e-06 ***
## anes2016$RaceBlack:2                0.3808063  0.3917677   0.972  0.33104    
## anes2016$RaceHisp:1                 0.3152484  0.3343170   0.943  0.34570    
## anes2016$RaceHisp:2                -0.1266669  0.3388402  -0.374  0.70853    
## anes2016$RaceNatAmAlsk:1            0.1085751  0.8237049   0.132  0.89513    
## anes2016$RaceNatAmAlsk:2            0.6106323  0.7697928   0.793  0.42764    
## anes2016$RaceOther:1               -0.4945788  0.4001184  -1.236  0.21643    
## anes2016$RaceOther:2               -0.1211030  0.3886383  -0.312  0.75534    
## anes2016$RaceWhite:1               -0.9843840  0.3045205  -3.233  0.00123 ** 
## anes2016$RaceWhite:2               -0.8309347  0.3055137  -2.720  0.00653 ** 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Names of linear predictors: log(mu[,1]/mu[,3]), log(mu[,2]/mu[,3])
## 
## Residual deviance: 6469.122 on 5918 degrees of freedom
## 
## Log-likelihood: -3234.561 on 5918 degrees of freedom
## 
## Number of Fisher scoring iterations: 5 
## 
## No Hauck-Donner effect found in any of the estimates
## 
## 
## Reference group is level  3  of the response

Results:

Interpretation of Highest Education Level -

Based on this analysis,

In comparison to voters whose highest level of education is an associate’s degree of some kind:

those whose highest level of education is a bachelor’s degree were less likely to identify as Democracts and even less likely to identify as Republicans, than those whose highest level of education is an associate’s degree of some kind.
those whose highest level of education is a high school diploma or GED were more likely to identify as Democracts and even more likely to identify as Republicans, than those whose highest level of education is an associate’s degree of some kind.
those whose highest level of education is less than a high school diploma or GED were more likely to identify as Democracts and * significantly more likely to identify as Republicans, than those whose highest level of education is an associate’s degree of some kind.
those whose highest level of education is a master’s degree were *** significantly more likely to identify as Democracts and less *significantly more likely to identify as Republicans, than those whose highest level of education is an associate’s degree of some kind.

Interpretation of Self-Identified Race -