Visualizing Associations Between Categorical (Factor) Variables

• Let's test and visualize association between race and protest behavior.
• Discuss re-ordering factors and changing level names

* Use mosaic plots to visualize association between factors

Read in the data

setwd("~/Dropbox/Data General/GSS")  #Set your working directory to whatever folder holds GSS.csv
options(scipen = 999)  #Turn off scientific notation
x <- read.csv("GSS.csv")  #Read in the dataset

table(x$attrally, x$race)  #View simple contingency table

##                                         
##                                          black other white
##   have done it in the more distant past     29    13   251
##   have done it in the past yr               22     8   153
##   have not done it and would never do it    84    37   285
##   have not done it but might do it          81    38   466

Levels of x$attrally are in weird order, let's re-order them...

Re-order Factor Levels

x$attrally <- factor(x$attrally, levels = c("have not done it and would never do it", 
    "have not done it but might do it", "have done it in the more distant past", 
    "have done it in the past yr"))

x$attrally <- factor(x$attrally, labels = c("0", "1", "2", "3"))

table(x$attrally, x$race)  #Review simple contingency table

##    
##     black other white
##   0    84    37   285
##   1    81    38   466
##   2    29    13   251
##   3    22     8   153

tab <- table(x$attrally, x$race)
chisq.test(tab)

## 
##  Pearson's Chi-squared test
## 
## data:  tab 
## X-squared = 29.23, df = 6, p-value = 0.00005505

mosaicplot(tab, color = TRUE, main = "Association Between Race and Protest")