After watching the Science and NASA channles in high def it got me thinking. I am guessing it has something to do with ground radar and maybe gps. Is it just a guess? How accurate can they get?
My guess is that they know how fast the Shuttle has to be going in the particular orbit that it’s in, and just quote that.
The Shuttle does use GPS data, now at least. I don’t think the GPS system was in place when the Shuttle program started, at which point it would be all from ground stations.
But it’s not hard to calculate, anyway. It’s only rocket science.
He shoots… he scores!
Various spacecraft use GPS, at least when they are close enough to Earth to be served by it.
I hate to nitpick, but it’s older than rocket science. It’s the sort of orbital mechanics that goes back to guys like Kepler and Newton, who knew nothing at all about rocket science.

I hate to nitpick, but it’s older than rocket science. It’s the sort of orbital mechanics that goes back to guys like Kepler and Newton, who knew nothing at all about rocket science.
And those guys didn’t even have digital computers, Matlab, and ground tracking stations. Heck, they actually had to use logarithms to make their calculations shorter. Logarithms! Heck, I haven’t even looked at a log table in over a decade.
Seriously, with some very simple optical tracking using a crude theodolite from a fixed position, a basic notion of the rotation of the Earth, and some high school trigonometry and analytical geometry, you could calculate the unpowered orbit (and thus, its position and velocity at any given time) of the Shuttle or any other observable object to three or four decimal places. It truly is, as Chronos says, not hard to calculate.
“Rocket science” (or “gas expansion propulsion science”) however–encompassing thermodynamics, reaction chemistry, avionics and nonlinear response control systems, compressible thermofluid dynamics, vibration and structural mechanics, to name a few disciplines–is actually pretty of hard stuff, sufficiently so that Kepler and Newton didn’t even have the tools to work in it, although Newton’s development of applied differential calculus (building upon Barrow’s fundamental theorem of calculus and properties of the tangent line) which permitted the underlying mathematical formalization of other sciences listed above.
Anyway, irrelevant and off-the-cuff history of science lesson aside, once you’ve measured the position of the Shuttle in a few places in its orbit (either from ground tracking, GPS measurements, or differential tracking from the Shuttle itself) you can make a very credible estimate of the orbital characteristics. From there, it’s all straight Newtonian mechanics.
Stranger
The Shuttle also has an inertial guidance system that keeps track of its position and velocity.
My understanding is that satellite orbits are calculated primarily from distance measurements made from the ground. It’s easy to shoot a radio signal at a satellite and either observe the reflection (i.e. radar), or receive a signal from a transponder on the satellite. From the time delay you can calculate the distance to the satellite very accurately. Do this continually as the satellite flies across the sky, and you get a set of distance vs. time data, from which you can calculate the orbit.
Measuring the position of the satellite in the sky is a little more difficult. Whatever instrument you use (telesccope, radio antenna, etc), you need high resolution, which usually means a narrow field of view. And you need to track the fast-moving object to keep it in the field of view.