source("http://www.uvm.edu/~rsingle/stat5210/data/scripts-5210.R")

#High school and beyond (hsb2) dataset: 200 obs
#scores on standardized tests of reading (read), writing (write), mathematics (math) and social studies (socst).
#demographic info: gender (female), socio-economic status (ses) ethnic background (race)
 dat <- otherdata("hsb2.dat")

 boxplot(write ~ ses, data=dat)
 mod <- lm(write ~ ses, data=dat)
 anova(mod) 
 class(dat$ses)
 
#re-code SES as a character variable
 dat$ses2 <- NULL
 dat$ses2[dat$ses==1] <- "low"
 dat$ses2[dat$ses==2] <- "med"
 dat$ses2[dat$ses==3] <- "high"
 
 boxplot(write ~ ses2, data=dat) #alphabetical variable ordering
 
#re-code SES as a factor
 dat$ses2 <- factor(dat$ses2, levels=c("low","med","high"))
 class(dat$ses2)
 
 boxplot(write ~ ses2, data=dat) #better variable ordering
 
 mod2 <- lm(write ~ ses2, data=dat)
 anova(mod2) 

#check assumption of equal variances 
 by(dat$write, dat$ses2, var)     #by(DV, IV, FUNCTION)
 tapply(dat$write, dat$ses2, var) #an alternative to by()

      var.test(write ~ ses2, data=dat)
 bartlett.test(write ~ ses2, data=dat)