Logistic regression!!
Ten little variables in my logistic line
Killed the one insignificant, now there are nine
Still insignificance seemed to be the model's fate
So I took off one more, now there are eight
Wald's scores told me the model was not even
One more variable died, and now there are seven
Poor chi-square scores put me in a terrible fix
Another independent killed, and now there are six
Now I felt, why at all should I keep them alive
Erase off one more, and now there are five
All this I was doing for a good concordant score
So another chap took the hit, now there are four
Confusion hits now, this killing like a decision tree
Oh what the hell!! One more down, now there are three
A finished model I desire now without much ado
So take off one more, and now there are two
Whee, removing independents is so much fun
I'll remove one more, now there is just one
Now I wonder, am I analyst or revenge-seeking hero
Why oh why did I bring the number down to zero!!