The way I had my students go about gearing this motor was pretty simple. (The details below are from memory, so they may be a bit off.)
This group wanted to make a robot fast enough to pop a wheelie. They identified acceleration, mass, and the location of center-of-mass as the key factors to achieve this.
I knew that a skid-steer robot would be hard to control at those speeds, and we thought having 1 motor near the back would help meet the weight & C.M. goals. That is why we chose the car-style steering and drive layout.
Since we didn’t have a spec sheet for the motor, we don’t know how much torque it has, so we started with the no-load rating of 4400RPM.
We then assumed that the motor would be most efficient at around half of this speed, 2200RPM.
We then figured out how far the 4" drive wheels would travel in 1 revolution (about 1 foot).
We then guestimated a top speed goal of around 4 ft per second. (This was a young team that was not quite up to the math & physics needed to calculate this properly, though we did white-board the forces at work so they could visualize it.)
That means the wheels needed to spin at a top speed of about 4RPS, or 240RPM.
So, we needed gearing to get us from 2200RPM down to 240RPM, which is a reduction of about 9:1.
I don’t recall the exact gear configuration, or the ratio we ended up with, but I believe we had a 2-stage reducer.
After some experimentation, we decided to switch out the 4" drive wheels for dual 2.75" wheels to slow it down a bit and provide better grip. It was nearly impossible to drive indoors… Oh, and they did meet their goal of popping a wheelie 
I could probably post a close-up picture of the motor shaft output this evening if that would be helpful for you.
Cheers,