#scripts & functions for STAT 221
 source("http://www.uvm.edu/~rsingle/stat221/data/scripts-221.R")

#lbs. of butterfat produced by each cow during a 305 day milking period at 2 different farms.

 dat <- otherdata("butterfat.dat")

 dat
 
 summary(dat)
 
 t.test(dat$X)

 t.test(dat$Y)

 boxplot(dat$X, dat$Y)

 t.test(dat$X, dat$Y)

 help(t.test)

 t.test(dat$X, dat$Y, var.equal=TRUE)

 t.test(dat$X, dat$Y, alternative="less")


#alternate data layout

 dat2 <- otherdata("butterfat2.dat")
 
 dat2

 t.test(dat2$pounds~dat2$farm)

 t.test(pounds~farm, data=dat2)

 boxplot(pounds~farm, data=dat2)



















#missing values 
 mean(dat$Y)             #returns the mean (705.4375)
 mean(dat$X)             #returns NA due to missing values
 mean(dat$X, na.rm=TRUE) #returns the mean after removing missing values (na.rm)
 
 sd(dat$Y)             #returns the Standard Deviation (SD)
 sd(dat$X)             #returns NA due to missing values
 sd(dat$X, na.rm=TRUE) #returns the SD after removing missing values

#Note that missing values are handled differently by different functions
 summary(dat)  #NAs are accounted for
 t.test(dat$X) #NAs are accounted for