"R provides Type I sequential SS, not the default Type III marginal SS reported by SAS and SPSS. In a nonorthogonal design with more than one term on the right hand side of the equation order will matter (i.e., A+B and B+A will produce different results)! We will need use the drop1( ) function to produce the familiar Type III results."
I'm a little disappointed in this article. It's a little incoherent and doesn't do a great job of talking about what it means by statistics. It's something about the political power of data collection?
Retroactive extensions! Those will sure encourage innovation. So shameful.
Completely bullshit/nonsense lecture that a bunch of professors completely buy and even rate highly! Delivery matters. Content, not so much.
More from Erich Goode on sex with his informants... Good God.
The most incredible piece of journalism I've read in a very long time. I just donated to Mother Jones.
step-by-step on how to mount an encrypted and LVMed disk image.
I love this kind of information archeology...
Lucas Brunelle jumps a pothole on his bike while be interviewed by people in a car. :)