data <- read.csv("~/glucinsul.csv")
head(data)

data <- data.frame(Time=c(data$Time,data$Time),
		cond=c(rep(0,50),rep(1,50)),gluc=c(data$basegluc,data$ofgluc))

# compute mean for each time and conditon combination
sum.tab <- tapply(data$gluc,list(data$Time,data$cond),mean)
#
# If you want to be an R pro, learn to use "plyr" library!
#

# make a plot
plot(data$Time,data$gluc,xlab="Minutes",ylab="Glucose",
	sub="red = Control, blue = Treatment")
points(c(0,30,60,90,120),sum.tab[,1],pch=16,col="red",cex=2)
points(c(0,30,60,90,120),sum.tab[,2],pch=16,col="blue",cex=2)

# make time a categorical variable
data$Time <- as.factor(data$Time)

# fit reduced model
fit.reduced <- lm(gluc~Time,data=data)
summary(fit.reduced)

# fit full model
fit.full <- lm(gluc~Time*cond,data=data)
summary(fit.full)

#compare the two
anova(fit.reduced,fit.full)

# to see where F came from:

((81.830 - 78.912)/5) / (78.912/90)

# NOTE: R calls sum of squared errors "RSS" (residual sum of squares)!!!