Lyric copyright by Dorry Segev
may sing to the tune of "It's Still Rock'N'Roll To Me" (Billy Joel)
What's the matter with the p I'm getting?
Can't you tell your CI's too wide?
Maybe I should check the line I'm fitting.
Type predict and don't ask why.
How have you been exploring all your data?
Things are automated if you do it all in STATA.
Everybody's talkin' 'bout regression
Funny, but it's just STATA code to me
What's the matter with my Kaplan-Meier?
Your control group is far too sick.
How'm I gonna make my hazard higher?
Use the backward stepwise trick.
Nowadays you can't be too automatic
You think you want a spline but it knows you want quadratic.
2-tail, no fail, do it on the log scale,
It's just STATA code to me
Oh, it doesn't matter what you write in your papers
Cause no-one has the sense to gauge
A forced model fills your needs
And JAMA never reads what you write in your methods page
We live in a peer-review age
Calculus doesn't lend itself to clarity
But we don't need to worry a bit.
Cause vif will check for collinearity
And estat, for goodness-of-fit.
When I'm feeling despair, worried and weary,
I remember that software is easier than theory.
Use 2 strata, you'll get a better beta,
It's just STATA code to me.
Fitting models now is my obsession,
My friends say I'm out of touch.
When I'm alone I like to do my own regression.
My friends say I think too much.
Don't you know about the glm command?
Gaussian, binomial, or logit on demand!
Use 2 strata, you'll get a better beta,
It's just STATA code to me.
Everybody's talkin' 'bout r-squared
Funny, but it's just STATA code to me