[answered] library(car) library(MASS) library(VGAM) library(effects) l

The following data set contains a dichotomous response variable, along?with other variables that can be treated as explanatory variables:?
library(car)

 

library(MASS)

 

library(VGAM)

 

library(effects)

 

library(nnet)

 

data(Chile)

 

C=Chile[complete.cases(Chile),]

 

C=C[C$vote!="A",]

 

C=C[C$vote!="U",]

 

vote_ind = 1*(C$vote=="Y")

 

m=lm(vote_ind~C$statusquo)

 

plot(C$statusquo, jitter(vote_ind, .2))

 

abline(m, col="red")

 

x=seq(-5,5,length=10001)

 

plot(x, pnorm(x), type="l")

 

lines(x, plogis(x, scale=sqrt(3)/(pi)), col="red")

 

data(Mroz)

 

m=glm(lfp~., data=Mroz, family=binomial(link=logit))

 

summary(m)

 

m2=glm(lfp~.-hc, data=Mroz, family=binomial(link=logit))

 

summary(m2)

 

m3=glm(lfp~.-hc-k618, data=Mroz, family=binomial(link=logit))

 

summary(m3)

 

anova(m2,m, test="Chisq") m0=glm(lfp~1, data=Mroz, family=binomial(link=logit))

 

summary(m0)

 

# Section 14.2.1 Polytomous Logit Model

 

data(BEPS)

 

BEPS$vote=factor(BEPS$vote, levels=c("Liberal

 

Democrat","Labour","Conservative")) #Reorder the levels to be consistent with

 

the book

 

m1=multinom(vote~.+Europe*political.knowledge, data=BEPS) #Fit the model

 

summary(m1) #Contains the same information as Table 14.5

 

Anova(m1) #Table 14.6

 

plot(effect('Europe*political.knowledge',m1,xlevels=list(Europe=seq(1,11,length=

 

50),political.knowledge=0:3)), style="stacked",rug=FALSE,

 

colors=c("gold","red","blue")) #stacked effects plot

 

plot(effect('gender',m1), style="stacked",rug=FALSE,

 

colors=c("gold","red","blue")) #stacked effects plot

 

plot(effect('age*economic.cond.national*gender',m1), style="stacked",rug=FALSE,

 

colors=c("gold","red","blue")) #stacked effects plot predict(m1, newdata=BEPS[1,2:10], type="probs") m2 = update(m1,.~.-gender) #Removing Gender from the Model

 

summary(m2)

 

1-pchisq(m2$deviance-m1$deviance,df=(m1$edf-m2$edf)) #testing whether the full

 

model m is significant improvement over the reduced model m2, it's not.

 

Anova(m2)

 

anova(m2,m1)

 

plot(effect('Europe*political.knowledge',m2,xlevels=list(Europe=seq(1,11,length=

 

50),political.knowledge=0:3)), style="stacked",rug=FALSE,

 

colors=c("gold","red","blue"))

 

m3=update(m2,.~.-economic.cond.household)

 

summary(m3)

 

1-pchisq(m3$deviance-m2$deviance,df=(m2$edf-m3$edf)) #testing whether the full

 

model m is significant improvement over the reduced model m2, it's not.

 

Anova(m3)

 

anova(m3,m2)

 

plot(effect('Europe*political.knowledge',m3,xlevels=list(Europe=seq(1,11,length=

 

50),political.knowledge=0:3)), style="stacked",rug=FALSE,

 

colors=c("gold","red","blue"))

 

plot(effect('Europe*political.knowledge',m3,xlevels=list(Europe=seq(1,11,length=

 

50),political.knowledge=0:3)),rug=FALSE) #same plot but 'unstacked' so you get

 

each probability curve seperately, with confidence bounds.

 

# Section 14.2.2 Nested Dichotomies

 

BEPS$vote.lab=with(BEPS, recode(vote, "'Labour' = 'yes'; else = 'no'"))

 

BEPS$vote.con=with(BEPS, recode(vote, "'Conservative' = 'yes'; 'Liberal

 

Democrat' = 'no'; 'Labour' = NA"))

 

m.lab=glm(vote.lab~age+economic.cond.national+economic.cond.household+Blair+Hagu

 

e+Kennedy+Europe+political.knowledge+gender+Europe*political.knowledge,

 

data=BEPS, family=binomial(link=logit))

 

m.con=glm(vote.con~age+economic.cond.national+economic.cond.household+Blair+Hagu

 

e+Kennedy+Europe+political.knowledge+gender+Europe*political.knowledge,

 

data=BEPS, family=binomial(link=logit))

 

predict(m.lab, newdata=BEPS[1,2:10], type="response")

 

predict(m.con, newdata=BEPS[1,2:10], type="response")

 

hel

 

# Section 14.2.3 Ordered Logit Model

 

data(WVS)

 

m1=polr(poverty~.+country*religion+country*degree+country*age, data=WVS,

 

method='logistic')

 

m2=vglm(poverty~.+country*religion+country*degree+country*age, data=WVS,

 

family=cumulative(parallel=TRUE)) #Section 14.3 Discrete Explanatory Variables

 

closeness=factor(rep(c("one.sided", "close"),c(3,3)),

 

levels=c("one.sided","close"))

 

preference=factor(rep(c("weak","medium",

 

"strong"),2),levels=c("weak","medium","strong"))

 

voted=c(91,121,64,214,284,201)

 

did.not.vote=c(39,49,24,87,76,25)

 

logit.turnout=log(voted/did.not.vote)

 

Campbell=data.frame(closeness,preference,voted,did.not.vote,logit=logit.turnout)

 

oldpar=par(mar=c(5.1,4.1,4.1,4.1))

 

with(Campbell, interaction.plot(preference, closeness, logit, type="b",

 

pch=c(1,16), cex=2, ylab="log(Voted/Did Not Vote)"))

 

probabilityAxis(side="right", at=seq(0.7, 0.875, by=0.025),

 

axis.title="Proportion(Voted)")

 

par(oldpar) m1=glm(cbind(voted,did.not.vote)~closeness*preference, data=Campbell,

 

family=binomial(link=logit))

 

m2=update(m1, ~.-closeness:preference)

 

summary(m1)

 

summary(m2)

 

anova(m2,m1)

 

Campbell.binary=data.frame(close=NULL, prefer=NULL, turn=NULL)

 

for (j in 1:6){

 

x1=with(Campbell, data.frame(close=closeness[j], prefer=preference[j],

 

vote=rep("did not vote", did.not.vote[j])))

 

x2=with(Campbell, data.frame(close=closeness[j], prefer=preference[j],

 

vote=rep("voted", voted[j])))

 

Campbell.binary=rbind(Campbell.binary, x1, x2)}

 

ftable(xtabs(~close+prefer+vote,data=Campbell.binary))

 

m1a=glm(vote~close*prefer, data=Campbell.binary, family=binomial(link=logit))

 

m2a=update(m1a, ~.-close:prefer)

 

anova(m2a, m1a)

 

---

Solution details:

About this Question

STATUS

Answered

QUALITY

Approved

DATE ANSWERED

Sep 18, 2020

EXPERT

Tutor

ANSWER RATING

GET INSTANT HELP/h4>

We have top-notch tutors who can do your essay/homework for you at a reasonable cost and then you can simply use that essay as a template to build your own arguments.

You can also use these solutions:

• As a reference for in-depth understanding of the subject.
• As a source of ideas / reasoning for your own research (if properly referenced)
• For editing and paraphrasing (check your institution's definition of plagiarism and recommended paraphrase).

This we believe is a better way of understanding a problem and makes use of the efficiency of time of the student.

NEW ASSIGNMENT HELP?

Order New Solution. Quick Turnaround

Click on the button below in order to Order for a New, Original and High-Quality Essay Solutions. New orders are original solutions and precise to your writing instruction requirements. Place a New Order using the button below.

WE GUARANTEE, THAT YOUR PAPER WILL BE WRITTEN FROM SCRATCH AND WITHIN YOUR SET DEADLINE.