Question1

GrabData.R

@@ -12,4 +12,5 @@ mydata3 <- filter(mydata2,BIGBANG=="True" | BIGBANG=="False",EVOLVED=="True"|EVO
 GSSdata <- filter(mydata3,CAPPUN=="FAVOR"|CAPPUN=="OPPOSE",VOTE12=="Voted"|VOTE12=="Did not vote",VOTE16=="Voted"|VOTE16=="Did not vote") %>% droplevels()
 levels(GSSdata$VOTE12)[1] <- "voted12"
 levels(GSSdata$VOTE12)[2] <- "no in 12"
-rm(Gss,Gss1,mydata,mydata2,mydata3)
+rm(Gss,Gss1,mydata,mydata2,mydata3)
+

gss2018.rmd

@@ -1,7 +1,7 @@
 ---
 title: "General Social Survey"
-author: "Your Name"
-date: "Year 2020"
+author: "Kaylie Brehm"
+date: "Summer 2022"
 output: 
   html_document:
     number_sections: true
@@ -26,29 +26,68 @@ source("GrabData.R")
 The data in the dataframe GSSdata is from the 2018 General Social Survey. The first blocks of R-code has selected down a subset of the data to just 16 variables. It has further removed unwanted factor levels in much of the data. Examine the code in the GrabData.R file to see what it is doing. Some of the variables are categorical and others are numerical. Be sure to do a variable analysis before tackling each question.  
 First question - Is opinion on the death penalty (CAPPUN) independent of gun ownership (OWNGUN)?
 
+$H_0$ Opinion on death penalty is not independent on ownership on gun.  
+$H_A$ Opinion on death penalty is independent on ownership on gun. 
 
 ## Methods
 
 <!--Decide on your methods:  use "variable analysis" or other appropriate descriptors.  Make sure to choose at least one graphical method and at least one numerical method.!-->
 
-##Results
+Both are categorical variables, each with two levels. Owning a gun would be a yes or no. Opinion on death penalty would be for or against. The analysis technique we will use is CAT~CAT. The results will show a bar chart, some numerical values, a fisher exact test for odds, and a chi-square test of independence.
+
+
+## Results
 
 <!--Divide this section into two sub-sections:  One for your descriptive  results and one for your inferential results.!-->
 
 ### Descriptive Results
 
+
+
 #### Graphical Descriptive Results
 
 <!--Graphical results here.  Make sure to show your code.  Provide appropriate labels for axes, giving units if possible, and provide a good title for the graph, too.  Use the graphical results to describe the patterns if any that exist in the data as focused toward the research question!-->
 
+We create two bar charts - one based on frequency and the other on percent.
+
+
+````{r}
+dd2 <- GSSdata %>% group_by(CAPPUN,OWNGUN) %>% summarize(count=n()) %>% mutate(prcnt=count/sum(count))
+# the group_by followed by summarize(count=n())
+basicC <- ggplot(dd2,aes(x=CAPPUN,y=count,fill=OWNGUN))
+basicC + geom_bar(stat="identity",position="dodge")
+#Now for percentage plot
+basicCC <- ggplot(dd2,aes(x=CAPPUN,y=prcnt*100,fill=OWNGUN)) 
+basicCC + geom_bar(stat="identity", position = "dodge")
+```
+
+Based on the data, it is apparent that those who oppose capital punishment, are more likely to say no to gun ownership. In those that favor capital punishment, slightly more people say no to gun ownership.
+
 #### Numerical Descriptive Results
 
 <!--Numerical results go here. Use the numerical results to describe the patterns if any that exist in the data as focused toward the research question!-->
 
+```{r}
+table2 <- xtabs(~CAPPUN + OWNGUN, data=GSSdata)
+rowPerc(table2)
+colPerc(table2)
+```
+
+The top data set shows percentages for each opinion on capital punishment in relation to opinion on gun ownership. About 70.97% of those who oppose capital punishment are against owning a gun. About 51.72% of those who favor capital punishment are also against owning a gun.
+
 ### Inferential Results
 
 <!--State hypothesis clearly.  Make sure your discussion of the inferential test covers all the aspects that the test output produces, such as test statistic, p-value etc.  Make a decision about the null hypothesis, explain the assumptions on which the selected test/procedure was based, and why the chosen procedure satisfys the assumptions and is appropriate to answer the research question!-->
 
+```{r}
+chisq.test(table2)
+chisqtestGC(table2)
+fisher.test(table2)
+```
+
+If the opinion for capital punishment is dependent on opinion on gun ownership, then there is a difference, meaning it is not 50/50 equal results. The Chi-Square adds up this difference and subtracts what we would expect if the null hypothesis were true. The P-Value is the probability that the null hypothesis is true. The null hypothesis was "Opinion on death penalty is not independent on ownership on gun." The p-value of the chi square test is 0.02022. Since this p-value is under 0.05, I reject the null hypothesis due to it being so small. The p-value of the fisher exact test is 0.01651. Since this p-value is under 0.05, I once again reject the null hypothesis due to it being so small. The odds ratio was 2.271 which is 1.271 away from one. So the probability that capital punishment opinion is dependent on gun ownership opinion is 127%.
+
+
 #  Question 2
 
 <!--In this section you explain what you are trying to show.  Where did the data come from?  What is the research or other question you are trying to answer?!-->