A scientific look at fuel economy

Well, I finally made a smart automotive decision.... I think. Since I could drive, I've always been in love with fast, large-engined cars. This flies in the face of being a science teacher and always telling people who whine about oil prices, "It's only going to get worse!" I've been calling for $5/gallon gas prices to come way before Al Gore ever had an inkling to make a reactionary powerpoint cut-and paste show. But I digress.
I'm putting the mustang in the garage except for play now; I would make the sensible move and sell it like my parents did with theirs, but there are three things holding me back:
1.) It's not worth that much money now.
2.) I was mad when I became a teen and found out my parents got rid of their mustangs!
3.) My lovely wife told me that I can't get rid of the car; turns out that's why she likes me after all!

I just picked up a used 1993 Ford Escort station wagon (I swore I'd never drive a wagon or anything the wagon became - Minivan or SUV, etc. when I was 16) . It has some miles on it (246k) and some little problems here and there. But it cost me $500. From my wicked mathematical abilities, I can calculate:
At today's gas price of $3.48, (Yes, it will go up another $1 within 2 years I think)
I spend $6.62 each day to get to work in my mustang, just on gasoline; 21 mpg.
Assuming a modest 31 mpg in the escort, that number is $4.49.
$2.13 savings per day, plus I get to keep the mustang garaged for the nice sunny days.
Payoff time for the car: 235 work days - a little over a year, if I just drive the car to work and don't use it for trips out of town. Any time the car runs past the 235 day mark helps my pocketbook, minus repairs - there will be some for sure.
Oh yeah, and each time gas prices go up, my payoff period shrinks. Something for new car buyers to think about.

Now, I don't really want prices to increase, so how can I get the car to be "worth it" in a shorter period of time?
Learning to drive better will be a start. I lurk on the tdiclub.com forums quite often, and have learned quite a bit from the turbo-dieselers. I can't afford a tdi yet. Definitely one for my wife on her next car. Apparently the best trick to get better gas mileage is to get a dozen eggs and strap one to the accelerator pedal and the other on the brake pedal. The learning curve is rough, but when you learn to drive without breaking the eggs, you've mastered it. ;)
Jokes aside, I'm excited to experiment with driving habits to observe effects. But what if that isn't enough?
Customizations. Lots of them. I'm pretty flexible, and I might have to learn some bodywork. I don't think I'm as good as the guy on 1989geometro.com, but I'll try.

What makes me different from every other joe out there trying to squeeze his pennies tighter? I know science. I'm willing to go to incredible ends in order to get dependable data. The number one mistake people make when they're "experimenting" is that they change lots of things at once, and then point and say, "Look, this did it!" That's called having more than one independent variable, in science teacher talk, and it means you don't know jack about who caused what. I'm going to experiment, using dependable methods, and I'm going to report my data to share it. And you can count on me being trustworthy and reliable.