The way GPS is used for tectonic studies is very different than how most people use GPS. Most people use GPS to determine a position and the GPS reports coordinates of that position (lat, long, alt).

Doing tectonic studies with GPS is quite different. Geophysicists don’t care about the position of any one point. Instead they measure the baseline between two points (dx, dy, dz). They then measure the same baseline again some time later (on the order of years) to see how the baseline has changed. They’ll do this for a number of baselines that not only span the plate boundary but also on each side of the plate. Here is a gross example (I’m going to try to do this with ASCII graphics)…

Imagine a plate bounday with points A and B on each side and a baseline between them:

A —– B

Assuming some plate movment has occured, several years later you measure the same baseline (ignore the periods):

A
.. \
….\
…..\
……B

As you can see the baseline orientation changes. Now do this for a network of baselines. Quantifiying these baseline change is what allows one to determine speed and direction of the plate movment.

In the example above, a plate movement where the left side of the plate boundary is moving upwards relative to the right side of the boundary can be quantified. Again, this is a very gross and exaggerated example but I hope you get the idea.

Generally, the actual points that are used for the endpoints of each baseline will be some kind of stable monument like a survey monument (but they don’t need to be surveyed)

So why baselines? The primary reason is precision/accuracy. Though GPS is very precise (some will argue accurate), it can only get your position to several meters, at best. The plate tectonic movement is well within this noise (so you won’t be able to trace any tectonic movement by monitoring positions). However, baseline measurements are extremely precise, on the order of millimeters or centimeters (depending on the length of the baseline). This is quite useful in detecting plate tectonic movement. However, this comes at a price. You cannot use consumer GPS products (TomTom, Garmin, etc) to do this kind of work. You need survey grade receivers. They cost tens of thousands of dollars and you’ll need several receivers to do this kind of work efficiently.
References :
Master Degree in Geophysics (Columbia University) specializing in plate tectonics. I also develop GPS receivers for a living.

http://hugebook.net/?p=244
References :