company <- read.table( "company.dat", header=T )
company$steel <- factor( company$steel )
company$site <- factor( company$site )
company$type <- factor( company$type )

company.fit <- aov( score ~ steel * site * type, company )
summary(company.fit)