##############################
#
# Script: Solutions 2 (assignment02a.R)
#
##############################



####################
# Problem 02.1

fb <- read.csv("football1.csv", header=TRUE)  # Read in the data as fb
b12 <- fb$score[fb$conference=="Big 12"]      # Select the Big 12 scores
sec <- fb$score[fb$conference=="SEC"]         # Select the SEC scores
t.test(b12,sec, alternative="greater")        # Compare them

# There is a shortcut in simple cases:
fb <- read.csv("football1.csv", header=TRUE)
t.test(fb$score~fb$conference, alternative="greater")

# This can also be done in simple cases:
fb <- read.csv("football1.csv", header=TRUE)
t.test(score~conference, data=fb, alternative="greater")

# So, why was the alternative = "greater" necessary?



####################
# Problem 02.2
gdp <- read.csv("gdpcap.csv", header=TRUE)   # Read in the data
g.afr <- gdp$gdpcap[gdp$region=="Africa"]    # Select only African gdpcaps
g.east <- gdp$gdpcap[gdp$region=="Eastern"]  # Select only Eastern gdpcaps
t.test(g.afr,g.east, alternative="greater")  # Perform the t-test

# Why no shortcut options here? What is different about this 
# data set from the one above? Why is that difference important?

# To be a good researcher, you must know your data!



####################
# Problem 02.3
frat <- read.csv("http://courses.kvasaheim.com/stat40x3/data/fraternity.csv")
gpa.ao <- frat$gpa[frat$fraternity=="AO"]
gpa.xp <- frat$gpa[frat$fraternity=="XP"]
t.test(gpa.ao,gpa.xp)

# or
frat <- read.csv("http://courses.kvasaheim.com/stat40x3/data/fraternity.csv")
t.test(frat$gpa~frat$fraternity)

# or
frat <- read.csv("http://courses.kvasaheim.com/stat40x3/data/fraternity.csv")
t.test(gpa~fraternity, data=frat)

# Notice I am pulling this data off the website without saving it
# to my disk first. I do not suggest this, unless the data will be 
# updated frequently.

# Finally, the boxplot
boxplot(gpa~fraternity, data=frat, las=1, ylab="GPA", main="GPA Comparison", xlab="Fraternity")

# This is the typical "Structure F" of all graphing commands. It
# is most usable when there is a dataset involved instead of two
# separate variables that you want to plot.

##### Extra:

# The means are not plotted on the boxplots. If you want the 
# means plotted, you will have to add them yourself:
points(1,mean(frat$gpa[frat$fraternity=="AO"]), col=2, pch=16)
points(2,mean(frat$gpa[frat$fraternity=="XP"]), col=2, pch=16)

# Note that the plotting function is "points" and not "plot". The
# latter will erase your current plot and start a new one. The former
# will add to the previous. This is a typical "Structure P", the 
# alternaive format to graphing commands. The variables are included
# in x,y format. Thus, the first point is plotted at the position(1, mAO),
# where 'mAO' is the mean of the AO fraternity.