LR5: Making and Comparing Confidence Intervals

#Lab Report 5: Making and Comparing Confidence Intervals
library(ggplot2)
library(dplyr)
library(lsr)
library(descr)
library(Hmisc)
library('lehmansociology')
data(gss123)
options(scipen = 999)

#For this LR we will be looking at the
#relationship between a dichotomous nominal independent variable
#and an interval ratio dependent variable
#First, we have to look at our dichotomous variable and see
#what the two categories are. 

#WHICH DICHOTOMOUS VARIABLE ARE YOU USING?

#Look at the two cateogories in your frequency table
#and decide what you will call each of your mini
#datasets. Choose names that make sense. 

#Below, REPLACE maledata AND femaledata WITH THE NAMES OF YOUR TWO
#MINI DATASETS. REPLACE sex WITH THE NAME OF YOUR 
#DICHOTOMOUS VARIABLE. REPLACE "Male" and "Female"
#WITH THE TWO CATEGORIES OF YOUR FREQUENCY TABLE.
#REMEMBER TO INCLUDE THEM IN QUOTATION MARKS AND
#TO WRITE THEM EXACTLY AS THEY APPEAR (FOR
#EXAMPLE, IF THEY ARE CAPITALIZED IN THE FREQUENCY
#TABLE THEY NEED TO BE CAPITALIZED IN THE CODE).

maledata <- subset(gss123, sex == "Male")
femaledata <- subset(gss123, sex == "Female")

#Follow the code below for the rest of your
#LR but replace maledata and
#femaledata with the names of your two mini datasets
#and replace sei10 with the name of your 
#interval ratio variable.
#Make sure to change titles and labels, too!
#Fill in your answers below

#WHICH INTERVAL-RATIO VARIABLE ARE YOU USING?

#WHAT IS YOUR RESEARCH QUESTION?

#Summary statistics and a histogram for your dependent variable 
#for your first mini dataset

#CHANGE THIS CODE BY REPLACING maledata WITH THE NAME
#OF ONE OF YOUR TWO MINI DATASETS AND REPLACING
#sei10 WITH YOUR DEPENDENT VARIABLE
#AND BY EDITING THE TITLES AND LABELS FOR GRAPHS
#REMEMBER TO THINK ABOUT BINWIDTH AND WHETHER YOU WANT
#TO CHANGE IT TO SOMETHING BIGGER

summary(maledata$sei10, na.rm=TRUE)
sd(maledata$sei10, na.rm=TRUE)

ggplot(data = maledata, aes(x = sei10)) + 
  geom_histogram(color="blue", fill="pink", binwidth =1, 
                 aes(y=(..count../sum(..count..))*100)) + 
  ggtitle("Distribution of Males' Socioeconomic Index Score, GSS") + 
  labs(y="Percent", x="Socioeconomic Index Score") 

#INTERPRET THE SUMMARY STATISTICS AND HISTOGRAM

#Now we are getting summary statistics and a
#histogram for your second mini dataset. 

#CHANGE THIS CODE BY REPLACING femaledata WITH THE NAME
#OF YOUR OTHER MINI DATASET AND REPLACING
#sei10 WITH YOUR DEPENDENT VARIABLE
#AND BY EDITING THE TITLES AND LABELS FOR GRAPHS
#REMEMBER TO THINK ABOUT BINWIDTH AND WHETHER YOU WANT
#TO CHANGE IT TO SOMETHING BIGGER

summary(femaledata$sei10, na.rm=TRUE)
sd(femaledata$sei10, na.rm=TRUE)

ggplot(data = femaledata, aes(x = sei10)) + 
  geom_histogram(color="blue", fill="pink", binwidth =1, 
                 aes(y=(..count../sum(..count..))*100)) + 
  ggtitle("Distribution of Females' Socioeconomic Index Score, GSS") + 
  labs(y="Percent", x="Socioeconomic Index Score") 

#INTERPRET THE SUMMARY STATISTICS AND HISTOGRAM

#Comparing means in a bar graph using the gss123 dataset
#CHANGE THE INDEPENDENT VARIABLE (replace sex)
#CHANGE THE DEPENDENT VARIABLE (replace sei10) 
#CHANGE THE LABELS
ggplot(data=gss123) +   
  stat_summary(aes(x=sex, y=sei10), fun=mean, geom="bar") +   
  xlab("Sex") +   
  ylab("Mean Socioeconomic Index Score") +   
  theme(axis.text.x=element_text(angle=-45))

#ANSWER THIS: Compare the results you got for your two groups. 
#Which has a bigger mean? Which has more variability? 
#Based on the bar graph of means, 
#does it seem like the two categories of your 
#dichotomous variable
#differ in the dependent variable?

#USING INFERENTAL STATISTICS TO COMPARE GROUPS
#Let's compare the confidence intervals for the mean of our
#dependent variable for the two categories of our independent variable.
#REPLACE maledata AND femaledata WITH THE NAMES OF YOUR MINI DATASETS
#REPLACE sei10 WITH THE NAME OF YOUR DEPENDENT VARIABLE

#Comparing the CIs- 95%
ciMean(maledata$sei10, na.rm=TRUE, conf =0.95)
ciMean(femaledata$sei10, na.rm=TRUE, conf =0.95)

#Comparing the CIs - 99%
ciMean(maledata$sei10, na.rm=TRUE, conf =0.99)
ciMean(femaledata$sei10, na.rm=TRUE, conf =0.99)

#INTERPRET the confidence intervals and come to a conclusion about whether
#your independent and dependent variables are related.