# analyzing data

rm(list=ls())
hs1 <- read.table("http://www.ats.ucla.edu/stat/R/notes/hs1.csv", header=T, sep=",")
attach(hs1)

# chi-square test of independence
tab1<-table(female, ses)
summary(tab1)

# one-sample test
t.test(write, mu=50)

# paired t-test
t.test(write, read, paired=TRUE)

# two-sample independent t-test
by(write, female, var)
tapply(write, female, var)
# assuming equal variances
t.test(write~female, var.equal=TRUE)  

# assuming unequal variances
t.test(write~female, var.equal=FALSE)  

# one-way anova
anova(lm(write~factor(prog)))
summary(aov(write~factor(prog)))

# an example of a two-way factorial ANOVA
# type I sum of squares
m2<-lm(write~factor(prog)*factor(female))
anova(m2)

# https://stat.ethz.ch/pipermail/r-help/2010-March/230280.html
# John Fox's post on type I, II and III sum of squares in ANOVA

library(car)
# type II sum of squares
Anova(m2)

# an example of ancova
anova(lm(write~factor(prog) + read))
summary(aov(write~factor(prog) + read))

# ols regression
summary(lm(write~female+read))

lm2 <- lm(write~read+socst)
summary(lm2)

# plotting diagnostic plots of lm2
plot(lm2)  

class(lm2)
names(lm2)
length(lm2)
length(lm2$residuals)
length(lm2$coefficients)
lm2$coefficients

# extracting predicted values and residuals
write[1:20]
fitted(lm2)[1:20]
resid(lm2)[1:20]


# logistic regression

# creating a binary variable
honcomp <- write >= 60
honcomp[1:20]

# fitting a logistic regression model
lr <- glm(honcomp~female+read, family=binomial)
summary(lr)

# odds ratios
exp(coef(lr))  

# Non-Parametric Tests

# analog to one-sample t-test
wilcox.test(write, mu=50)

#  analog to the paired t-test
wilcox.test(write, read, paired=T)

#  analog to the independent two-sample t-test. 
wilcox.test(write, female)

# anallog to one-way anova 
kruskal.test(write, ses)

detach(hs1)